Toothed whales (Order Cetacea; suborder Odontoceti), including dolphins and porpoises, are notable among mammals in having high levels of directional asymmetry (a deviation from symmetry in the same direction in the vast majority of individuals) in their skulls, particularly those elements associated with the airway (Ness, 1967; Yurick and Gaskin, 1988). Odontocetes are also notable as being one of the few mammalian lineages that have evolved the use of biosonar to navigate and forage (Moore, 1980; Thompson and Richardson, 1995). This finding has led to an assumption that these two characteristics of odontocete cetaceans, directional skull asymmetry and biosonar usage, must be linked and that skull asymmetry is an adaptation for the production and use of biosonar (e.g., Yurick and Gaskin, 1988). This linkage is supported by the numerous other soft tissue adaptations for biosonar production and sound transmission found in the head of odontocetes (e.g., Heyning, 1989; Cranford et al., 1996; Cranford and Amundin, 2004; Goodson et al., 2004). It would thus seem logical that skull asymmetry is also an adaptation for sound production and transmission, particularly because the asymmetry is primarily displayed in conjunction with soft tissue adaptations for biosonar.
However, odontocete species vary greatly in the extent of skull asymmetry (Ness, 1967), and it is unclear how this variation could relate to any functional link between asymmetry and the use of biosonar. For example, little work has been done to investigate if and how interspecific variations in skull asymmetry relate to differences in sound production, transmission, form, function, or prey preferences between odontocete species. It is also not clear whether skull asymmetry is even a requirement of, or beneficial to, the use of biosonar. Certainly, directional skull asymmetry is not a feature of the other major lineage that has evolved echolocation (microchiropteran bats), demonstrating that a functional echolocation system can evolve without associated skull asymmetry. However, there are several differences in the mechanism of sound production, the form of sounds used for echolocation, and the acoustic environment for bats versus odontocetes (see Cranford and Amundin, 2004), as well as other ecological characteristics, that may invalidate any direct comparisons between bats and cetaceans in relation to an acoustic function of skull asymmetry. Thus at present, skull asymmetry is assumed to be linked to biosonar production, despite that there are no hypotheses to link observed variations in skull asymmetry to functional variations in biosonar production or usage, no explanation of why skull asymmetry is beneficial for biosonar in odontocetes, and evidence that skull asymmetry is not, per se, a direct requirement for the development of biosonar.
The skull and nasal soft tissues are not the only directionally asymmetric characteristic within the heads of odontocete cetaceans. The larynx is also known to be directionally asymmetric and to deviate to the same side as many of the asymmetric elements of the skull (Reidenberg and Laitman, 1994). Rather than being linked to the use of biosonar, the functional reason for this asymmetry is thought to be primarily related to a conflict between the requirements of the airway and the digestive tract that arises because cetaceans must catch and ingest food while underwater (see Reidenberg and Laitman, 1994, for a full discussion of function of the asymmetrical positioning of the hyolaryngeal apparatus; see Werth, 2007, this issue, for discussion of hyolingual apparatus anatomy and role in feeding).
As in all mammals, the airway and the digestive tract of cetaceans intersect. Therefore, there is a risk of food passing into the airway or air into the digestive tract. Heyning and Mead (1996) noted that there is no muscular sphincter at the anterior esophagus and hypothesized that the forestomach acts as a reservoir where water ingested along with prey can be held until the prey item has been secured (see Mead, 2007, this issue, on cetacean stomach anatomy). Once the prey item has been secured, this water can then be ejected back along the esophagus and out through the oral cavity. As a result, cetaceans may face the added risk of water passing into the airway both when prey is ingested underwater and/or when swallowed water is regurgitated. This problem is particularly an issue in the marine environment, as even a small amount of sea water entering the lungs has the potential to be extremely dangerous. For example, in humans the inhalation of salt water can result in “secondary drowning” or acute respiratory distress syndrome (Sachdeva, 1999; Mortelliti and Manning, 2002). However, a mechanism has evolved in odontocete cetaceans that will prevent any water ingested along with prey from entering the lungs under normal conditions. This mechanism is the complete isolation of the airway and digestive tract. To achieve this, the larynx penetrates straight through the pharynx from the ventral to the dorsal side and is permanently held in place by a ring of muscle (palatopharyngeal sphincter) surrounding its rostral extension (see Reidenberg and Laitman, 1987, 1988, for full anatomical details of structure and placement of the larynx). The encircling palatopharyngeal sphincter prevents food or water, passing by means of the pharynx into esophagus, from entering into the airway through the larynx.
While this adaptation removes the risk of water accidentally entering the airway during ingestion, the larynx effectively forms a blockage in the digestive tract, limiting the size of items that can be swallowed without risk of them becoming lodged around the larynx. Extant cetaceans do not, in general, process their food by mastication before ingestion, and most prey are swallowed whole. The only species known to regularly process food in any way is the killer whale (Orcinus orca); it can rip large cetaceans and other marine mammals into smaller pieces before ingesting them (Dahlheim and Heyning, 1999). However, killer whales still swallow smaller prey whole, such as fish. This general lack of processing of prey means that the position of the larynx will effectively limit the maximum size of prey that can be swallowed.
While this restriction on prey size is not thought to be a problem for baleen whales (suborder Mysticeti), that consume very small prey relative to body size, it prevents odontocetes from swallowing any relatively large prey they catch. Any attempt to swallow a food item too large to pass around the larynx could result in the larynx becoming dislodged and/or the food item becoming wedged in the throat. Either of these scenarios would be lethal, probably due to asphyxiation resulting from the dislodged larynx, and such fatalities are occasionally recorded in wild odontocete cetaceans (e.g., Siebert et al., 2001). As a result, odontocetes have evolved an asymmetrically placed larynx and unequal-sized piriform sinuses, thereby allowing them to swallow larger prey along the larger lateral food channel and reducing the risk of prey becoming lodged around the larynx with fatal consequences for the cetacean (Reidenberg and Laitman, 1994). Under these circumstances, species may vary in the level of laryngeal asymmetry, depending upon the selection pressure placed on them by their prey size preferences, with the asymmetry of each species being tailored to its own specific niche (particularly prey size aspects of this niche).
Building upon this, it is proposed that this adaptive directional asymmetry is not limited to the larynx, but rather is displayed in any section of the respiratory tract where there is no over-riding selection pressure to remain symmetrical, including certain elements of the skull. For example, the posterior choanae (the lower or inferior ends of the nares) do not usually develop asymmetrically. This finding is probably because such asymmetry would result in asymmetry of the pterygoid, palatine, and possibly even the maxillary bones, which would likely compromise the function of the upper jaw, buccal cavity, and upper esophagus in this area. However, the anterior choanae (the upper or superior ends of the two nares) have no similar selection pressure to be symmetrical. As a result, the anterior choanae are free to display any selection pressure on the upper airway to develop asymmetrically to allow larger prey items to be swallowed. If this is correct, skull asymmetry in odontocetes is not an adaptation per se, but rather a developmental by-product of the need to completely separate the airway and the digestive tract to prevent accidental drowning (particularly secondary drowning), while still being able to swallow relatively large prey whole without risk of them getting stuck in the throat around the larynx. In addition, while asymmetry of the skull, and possibly other elements of the upper respiratory tract, may contribute toward the functioning of cetacean biosonar, this hypothesis suggests that directional skull asymmetry did not evolve solely as an adaptation associated with biosonar. In fact, under this hypothesis, it is feasible that the evolution of directional skull asymmetry preceded biosonar evolution.
If this hypothesis is correct, we predict that the maximum prey size consumed by a species will be related to level of asymmetry of the anterior choanae, with proportionately more asymmetric species consuming proportionately larger maximum prey sizes. Second, species with a narrower gape would be expected to have more symmetrical skulls as any gape limitation will limit the size of prey that can be ingested and so will reduce the selective pressure on the upper airway to be asymmetrical.
MATERIALS AND METHODS
This study used data on prey size and skull asymmetry from 13 species of odontocete cetaceans, including four narrow-gaped genera (Mesoplodon, Ziphius, Hyperoodon, and Kogia) and eight wide-gaped genera (Phocoena, Delphinus, Stenella, Lagenorhynchus, Tursiops, Grampus, Globicephala, and Orcinus). The purpose of this study was to investigate whether the maximum size of prey consumed was related to skull asymmetry. Stomach contents were examined from 294 specimens, including 15 from narrow-gaped species and 279 from wide-gaped species, to assess prey size. All specimens derived from the northeast Atlantic Ocean. Skulls were examined from 183 specimens, including 34 with narrow gapes and 149 with wide gapes, to assess asymmetry of the anterior choanae. This geographic area was chosen due to the availability of both stomach contents information from stranded animals and skulls. In addition, by concentrating on a single region, it reduced the possibility that geographic variations in skull morphology would mask any relationships between local prey preferences and the levels of asymmetry in the skulls of local cetacean populations.
Stomach contents from stranded animals were used to identify the maximum prey size relative to body size recorded in a species in the northeast Atlantic. Between 1 and 177 stomachs were measured per species depending on availability (Table 1). These stomachs had been analyzed as part of an ongoing program looking at dietary preferences in cetaceans (e.g., Santos et al., 1999, 2001a, b, 2004). From these data, a maximum predator–prey size ratio (PPSR) was calculated for each individual cetacean by dividing the size of the prey item with the largest size (estimated from hard remains such as otoliths or squid beaks using standard regression equations; fish: Härkönen, 1986; cephalopods: Clarke, 1986) by the body length of that individual cetacean. This strategy standardized the prey size in relation to the individual that contained it and allowed a comparison to be made between the maximum size of prey consumed by different individuals. The greatest PPSR recorded in each species was taken as indicative of the largest size of prey, relative to body size, that a species consumes in this region.
Table 1. Sample sizes and measurements for average skull asymmetry of the anterior choanae and maximum relative prey size in cetaceans from the northeast Atlantic Ocean
Genera species, common name
No. of skulls measured
Average relative asymmetry of the anterior choanae
No. of stomachs examined
Maximum relative prey size recorded
Mesoplodon bidens, Sowerby's beaked whale
Ziphius cavirostris, Cuvier's beaked whale
Hyperoodon ampullatus, Northern bottlenose whale
Kogia breviceps, Pygmy sperm whale
Phocoena phocoena, Harbour porpoise
Delphinus delphis, Common dolphin
Stenella coeruleoalba, Striped dolphin
Lagenorhynchus acutus, White-sided dolphin
Lagenorhynchus albirostris, White-beaked dolphin
Tursiops truncates, Bottlenose dolphin
Grampus griseus, Risso's dolphin
Globicephala melaena, Long-finned pilot whale
Orcinus orca, Killer whale
The asymmetry of the anterior choanae was measured in skulls held at the National Museums of Scotland in Edinburgh, United Kingdom. Between 2 and 24 skulls were measured per species depending on availability (Table 1). These skulls came from animals stranded in the geographic area for which stomach contents data were available, providing a consistent geographic location for the two data sets. The difference in width of the two anterior choanae was calculated. To account for size differences in skulls within and between species, and allow a comparison of the levels of asymmetry, the difference in width of the anterior choanae was standardized by a measure of skull size. In many studies of skull morphometrics, the skull length, measured from the tip of the rostrum to the back of the skull, is used to standardize other measurements. However, with museum specimens, particularly for species such as beaked whales with relatively delicate rostral bones, it is often not possible to accurately measure this due to damage to the tip of the rostrum. Therefore, we standardized the asymmetry measurements using maximum skull width, a feature that could be accurately measured in more specimens and that is less prone to damage. This measurement was justified on the basis that skull width is strongly correlated with both skull length (R2 = 0.905; n = 242; P < 0.0001) and with body length (R2 = 0.859; n = 175; P < 0.0001) in odontocete cetaceans (Weller, 2004). Therefore, by standardizing the difference in width of the two anterior choanae by the skull width, it was possible to produce a size-independent measure of the level of asymmetry using the following formula: (WRAC − WLAC)/WS, where WRAC is the width of the right anterior choanae, WLAC is the width of the left anterior choanae, and WS is the maximum width of the skull.
An average relative asymmetry value was calculated for each species. This strategy allowed the typical level of asymmetry in the upper airway within the skull for a species to be compared with the maximum relative prey size that each species is known to consume in the study area. The cetacean species were divided into two groups based on the extent of their gape. These were species with a narrow gape (beaked whales and sperm whales) and species with a wide gape (porpoises and dolphins). Beaked whales have a limited gape in comparison to delphinids (Heyning and Mead, 1996) and in comparison to porpoises. Sperm whales can similarly be considered to have a limited gape due to the relatively narrow under-slung mandible, and the narrow opening to the esophagus that results from this, in comparison to delphinids. This separation into two groups removed any potential effects of gape-limitation on the maximum size of prey that could be ingested and, therefore, the strength of the selective pressure on the larynx to be asymmetrical. Regression analysis was used to test whether there was a relationship between maximum relative prey size recorded from a species and the average relative asymmetry of the anterior choanae within the wide and narrow-gaped groups.
Within each group, there was a significant relationship between the asymmetry of the anterior choanae and the maximum prey size recorded from the stomachs (narrow gape: R2 = 0.909, regression analysis of variance [ANOVA] F = 31.01, P = 0.031, N = 4; wide gape: R2 = 0.642, regression ANOVA F = 15.32, P = 0.006, N = 9; Fig. 1). In both cases, a large proportion of the variation in maximum relative prey size could be explained by variation in the average asymmetry of the anterior choanae between species (64% and 91%, respectively). This finding is consistent with the predictions of the hypothesis. However, the relationship between maximum prey size and asymmetry of the anterior choanae differs between these two groups. The slope of the regression line was shallower for narrow-gaped species and in general, for any given level of asymmetry, the narrow-gaped species consumed a much smaller maximum relative prey size than wide-gaped species. This difference presumably reflects the effect that gape-limitation has on the size of prey that an odontocete species can consume.
Within the wide-gaped species, the killer whale displayed one of the lowest levels of relative asymmetry in the dorsal nares (Table 1), consistent with its ability to process large mammalian prey into smaller pieces before ingestion. However, all prey recorded from the single killer whale stomach used in this study were fish that had presumably been ingested whole. In this sample, the maximum relative prey size of prey recorded in the stomach was consistent with the general relationship between skull asymmetry and prey size. This finding suggests that, when consuming smaller prey items whole, killer whales are subject to the same constraints as other wide-gaped species.
The relationship between relative prey size and relative asymmetry of the anterior choanae is consistent with the hypothesis that skull asymmetry is linked to pressure to alter the position of the larynx to increase the maximum size of prey that can safely be consumed. Differences between species in levels of asymmetry presumably relate to niche differences in prey size preferences that lead to differing levels of selective pressure for asymmetry. In contrast, if skull asymmetry is linked to the use of biosonar, it would be difficult to explain the pattern of interspecific variation in asymmetry and the relationship with relative prey size, particularly when both have been standardized to remove any potential confounding effects of body size. While some aspect of biosonar linked to asymmetry could be related to the absolute size of prey a species consumes (for example, species that feed on small prey may be required to use higher frequencies to provide better resolution of small objects), this would not explain why asymmetry would be linked to relative prey size, where the influence of absolute prey size is removed.
Effective isolation of the airway and digestive tract is a requirement for any air-breathing aquatic mammal and frees animals from having to return to the surface to swallow any prey that are caught, making foraging more efficient. Therefore, it is likely that this was a key event in cetacean evolution and that it occurred at a relatively early point, allowing them to fully exploit the marine environment. A second key event in cetacean evolution was the shift from a heterodont dentition pattern, with some teeth adapted to process prey—allowing large items to be swallowed in smaller pieces, to a homodont dentition pattern, with all the teeth displaying the same general shape. While homodont dentition is better adapted for catching and holding aquatic prey, it does not allow food to be processed into smaller pieces, so requiring it to be swallowed whole. Homodont dentition is the general pattern in modern odontocetes, although there are some interesting derived variations in certain species (e.g., the unpaired tusk of the male narwhal; presence of only mandibular teeth in several species, such as sperm whales; the single pair of unusually shaped mandibular teeth of male beaked whales, while females do not erupt any teeth). Homodont dentition is also recorded in at least some extinct species of both odontocetes and mysticetes (e.g., Aetiocetus polydentatus; Barnes et al., 1995). Extant mysticetes, however, are currently toothless and rely on baleen plates for filter feeding. Therefore, homodont dentition may have evolved before the separation of these two lineages.
Once homodont dentition evolved, prey could no longer be easily processed. As long as cetaceans relied on catching and swallowing relatively large prey whole, there would be selective pressure to shift the larynx to allow relatively larger prey to be swallowed and reduce the risk of a potentially fatal blockage occurring at this point in the digestive tract. If directional asymmetry in toothed whale skulls has evolved as a by-product of asymmetrical laryngeal positioning, then it is likely that it would have appeared in the evolution of cetaceans in conjunction with, or soon after, the evolution of homodont dentition. This finding means that directional asymmetry is likely to have evolved only once in the cetacean lineage, along with homodont dentition, potentially explaining why all known odontocete species deviate from symmetry in the same direction, as suggested by Heyning (1989), while still providing an explanation of why species differ in their levels of asymmetry (Milinkovitch, 1995).
In addition, as some early baleen whales were homodont, skull symmetry in baleen whales could be derived rather than ancestral (Milinkovitch, 1995), following a switch from consuming a small number of large single prey items to filter feeding on large numbers of very small prey items. This switch in foraging technique would remove any selective pressure to have the larynx positioned asymmetrically, allowing the ancestral lack of directional asymmetry (found in almost all mammals, including the closest terrestrial ancestors to cetaceans) to be re-established. Thus, in phylogenetic analyses within the order Cetacea, it should not be assumed that skull asymmetry is a derived characteristic, as skull asymmetry may have evolved in a common ancestor of both odontocetes and mysticetes. If skull asymmetry did evolve before these two lineages split, any asymmetry in a common ancestor would have deviated to the left, as in all extant odontocetes cetaceans. In addition, skull asymmetry may have preceded the evolution of echolocation in the derived odontocete lineage and should not be used as a marker for the existence of functional biosonar in fossil cetaceans.
However, it is also possible that, despite a homodont dentition in some lineages, early baleen whales did not evolve an asymmetrically placed larynx. Instead, when faced with the same problem of getting prey passed the larynx where it crosses the digestive tract, rather than develop an asymmetrical larynx, baleen whales shifted their foraging strategy to capturing large numbers of easy-to-swallow small prey at any one time. Under such circumstances, the problem of separating the airway and digestive tract could have resulted in a selective pressure toward filter-feeding, and laryngeal and associated skull asymmetry would indeed be a derived characteristic unique to the odontocete lineage. These possibilities could be tested by looking for evidence of skull asymmetry in homodont ancestors of baleen whales and by examining the structure and positioning of the larynx in modern baleen whales (see Reidenberg and Laitman, 2007a, this issue, and 2007b, this issue, for more information on the larynx of baleen whales).
Therefore, the prey size hypothesis for the evolution of skull asymmetry in cetaceans potentially provides a single parsimonious explanation for variations in skull asymmetry throughout the entire order of Cetacea, rather than treating the asymmetry commonly observed in odontocete skulls as a special case. In addition, it explains the variations in the level of asymmetry between species without the need to postulate repeated and independent evolution of skull asymmetry to the same (left) side in several separate odontocete lineages (Milinkovitch, 1995). It also suggests that there may be linkage between many aspects of the development of cetacean skulls, and particularly skull asymmetry, with evolutionary adaptations to one element resulting in changes in associated elements through shared developmental processes. Finally, the anatomical relationships identified in this study between oral gape size, asymmetry of the anterior choanae, asymmetry of laryngeal placement, asymmetry of pharyngeal widths, and prey size may prove useful in reconstructing the soft tissues of the throat of extinct cetaceans in which only fossil skulls are available. Determining the extent of asymmetrical positioning of the larynx and, thus, width of the piriform sinuses, and by extension the maximum prey size that could be consumed, would help illuminate extinct cetacean feeding ecology. The same principles can also be used to determine feeding behaviors of rare modern cetaceans for which only skeletal remains may be available.
The authors thank the National Museums of Scotland for providing access to skulls from their collections for measurements. Analysis of stomach contents material collected up to 1997 was funded by the Commission of the European Communities. Sample collection and analysis during 1998–2000 was funded under CEC DG XIV Study Project 98/089. Sample collection and analysis during 2001–2003 was funded under CEC project EVK3-CT-2000-00027. The authors thank H. Ross, R. Reid, and I.A.P Patterson (SAC Veterinary Science Division, Inverness, Scotland) for collecting the stomach samples from Scottish animals during postmortem examinations. Finally, the authors thank the two anonymous reviewers for their comments and suggestions on this manuscript.