Monday, November 3, 2014

The evolutionary history of walruses, part 5: what did tusks evolve for?

For earlier entries in this series, see:

Sexual dimorphism in the walrus - male in the background, female in the foreground. Photo from

Tusks in Odobenus rosmarus

Tusks in the modern walrus Odobenus rosmarus occur in both sexes, but are generally larger and longer in males – and like most other pinnipeds, they are polygynous (a single male mates with multiple females) and sexually dimorphic (males are larger than females). The walrus is restricted to the Arctic – and owing to this, tusks were usually assumed to have something to do with ice. For example, walruses tend to use their tusks to assist in hauling out onto ice, leading many to originally propose that tusks evolved for this purpose. Other workers erroneously identified tusks as being used for excavation of mollusks on the seafloor. However, observations by Francis Fay (1982) and Edward Miller (1975) indicate that a use in feeding or haul out behavior is unlikely. Miller (1975) studied aggressive behavior in male walruses, and observed that tusks perform a central role in male interactions. Most interactions consist of tusk threat displays – the aggressor leans his head back so that the tusks are horizontal and pointing toward the target. If the target is somewhat submissive the aggressor will perform a “stabbing” motion. In more aggressive interactions the aggressor strikes the target with the tusks using the same downward stabbing motion, typically striking the hindquarters, back, or neck. These strikes commonly draw blood but Miller (1975) doubted that many cause serious injury (similar to elephant seal combat). Tusks were also frequently used to parry strikes close to the face. Predictably, walruses preferentially threatened smaller males; perhaps more adorably, juvenile males that even lacked tusks performed play fighting that was similarly ritualized. Strikes tended to follow visual threats, and Miller (1975) indicated that ritualized aggressive behavior like this is fundamentally similar to that seen in sea lions, who perform visual displays (prior to striking) by similarly leaning back and opening the mouth to show the canines. Interestingly, the pattern of scarring is completely opposite to the pattern of observed tusk strikes: scarring is mostly present on the anterior neck region, and Miller (1975) attributed this to several reasons: 1) his observations were on land and 2) during the summer. He hypothesized that during the breeding season, more intense “face to face” combat on ice (or more likely, in the water, as some rare anecdotes suggest) is the origin of anterior scarring. So, the relatively violent behavior that Miller (1975) described is not even that which is known to cause the most scarring on walruses – which seems to suggest that walrus breeding behavior might be a bit terrifying and may give the elephant seal a run for its money.

Walrus tusk display and combat. Threat displays frequently prelude tusk strikes. Photo from

            Many earlier workers (see Fay, 1982: 134-135 and references therein) concluded that walruses dug prey items out of seafloor with their tusks, and this was based primarily on observations of tusk abrasion in dead animals. At least one early study suggested that walruses scraped the seafloor with their tusks in a posterior direction, but later revised to a side-to-side motion as no abrasion exists on the posterior side of the tusks. Some early reports did cast doubt upon these hypotheses, as occasional individuals were identified as lacking tusks but of otherwise healthy appearance. Fay’s (1982) classic study re-examined the abrasion patterns, and concluded that the primary direction of sediment-tusk interaction was from proximal to distal (e.g. base of tusk to tip), which indicates that tusks are passively dragged through the sediment during benthic foraging. Fay (1982) also indicated that tusks are frequently used for locomotion – including hauling out onto sea ice, and even during aquatic sleeping with the tusks hooked over the edge of an ice floe (like a swimmer resting at the edge of a pool). However, he suggested that these were secondary functions and that by far and away the most significant functions were all social in origin. He hypothesized that because all/most pinnipeds are polygynous, the capability for tusk development is probably universal among the group but extreme canine enlargement is probably only possible once a pinniped lineage has made the shift from piscivory (fish eating) to suction feeding. Notably, most toothed whales with tusks (beaked whales, narwhal, Odobenocetops) are all either known or inferred to be suction feeders.

 Abrasion of walrus tusks - figure from Fay (1982). Abrasion is focused on the anterior side of the tusk, indicating passive dragging of the tusks through sediment during foraging rather than active digging.

Temperate and Subtropical tusked walruses

Further eroding ice-related hypotheses for the evolution of tusks in walruses are discoveries of fossil walruses that inhabited drastically warmer waters than the extant Odobenus rosmarus. The earliest known temperate tusked walrus was Alachtherium, which for the past 130 years was known from Belgium, and in the late 1990’s was also reported from the northwestern coast of Africa (Geraads 1997). Subsequently, additional discoveries indicated more occurrences of Alachtherium from Japan and the eastern USA as far south as Florida, and records of the toothless odobenine walrus Valenictus from southern California and even Mexico (Deméré, 1994). Fossils of Valenictus from San Diego and the Imperial Desert indicated to Deméré (1994) that walrus tusks evolved long before walruses became ice-bound in the Arctic, and that tusks are thus “structures with history”.

 Life restoration of Odobenocetops by Smithsonian artist Mary Parrish.

The walrus-faced whale Odobenocetops: implications for tusk use

The 1990 discovery of a bizarre fossil mammal, named in 1993 by Christian de Muizon as Odobenocetops, led to a reinterpretation of tusk function in walruses. Odobenocetops was collected from late-Miocene strata of the Pisco Formation of Peru and initially accidentally misidentified as a walrus; I’ve been told that an early SVP abstract with this mistake can be found. LACM Curator Emeritus takes credit for setting the record straight and asking those involved “why does the skull have premaxillary sac fossae?” These fossae, for the uninitiated, are unique to odontocetes (toothed whales), and Muizon (1993) named it as a new genus and species in a new family, Odobenocetopsidae, which he and others (Muizon et al., 2002) considered to be a sister clade to the Monodontidae – the family that includes the beluga and narwhal (and the fossil belugas, Bohaskaia and Denebola). I won’t go into too much detail irrelevant to the tusks, but Odobenocetops only possesses two teeth: asymmetrical left and right tusks that are posteriorly directed and set into elongate, columnar alveolar processes, and exhibits a deeply concave palate. These features and their similarity with the modern walrus indicated a similar mode of feeding. However, the occurrence of similar tusks in a completely different type of marine mammal that independently evolved benthic suction feeding for mollusks begs the question: did tusks really evolve for social purposes? Muizon et al. (2002) conclude that the orientation of the tusks is a bit too coincidental, and that the alveolar processes likely behaved as “sled runners” to stabilize and properly orient the head of Odobenocetops as it trawled the ocean floor for molluscan prey. They conceded that the asymmetry of the tusks (the left tusk is barely erupted while the right tusk is very long – up to 1.35 meters in Odobenocetops leptodon; Muizon and Domning, 2002) indicates that such a function was not optimized in Odobenocetops, and it likely reflects a social function like the tusk of the narwhal.

Seafloor foraging of a walrus. From this paper by Levermann et al.

Speaking of tusked cetaceans… what the heck is the narwhal tusk for?

This is a bit of a convenient topic to tack on here; I’d like to revisit it in more detail in the future since some interesting papers have come out in recent years on the topic. The narwhal (Monodon monoceros) is also sexually dimorphic, and possesses a pair of tusks, generally only the left tusk erupts from the soft tissue. Rarely males will possess an erupted right tusk. Although formerly considered an incisor, recent CT studies indicate that the tusk is embedded entirely within the maxilla and is therefore the canine tooth; a series of other vestigial postcanine teeth also form (Nweeia et al. 2012) but rarely erupt from the skull or soft tissues (and are therefore detectable only using CT imaging). Sexual tusk dimorphism is a bit more extreme than in the walrus: only 15% of female narwhals ever possess tusks that erupt from the soft tissue, and the tusks are always smaller and shorter than those of males. Significantly, narwhals do not appear to be polygynous. The narwhal tusk is conspicuously “spiraled” (presumably for structural rigidity) and exhibits dentine tubules exposed on the surface of the tooth – which suggests some ability to sense water temperature and salinity (Nweeia et al. 2009). In contrast, in mammals that masticate their food the dentine tubules do not extend to the outer margin of the tooth; indeed, toothaches may be caused by dentine tubules being exposed to the oral environment when a cavity forms. Additionally, a pulp cavity extends along the entire length of the tusks. Field experiments which consisted of exposing a small section of tusk to high salinity solution resulted in rapid head movements and breathing in several different individuals. These observations lead Nweeia et al. (2009) to propose that the narwhal tusk fulfills a sensory function. 

Male and female narwhals underwater. There are surprisingly few underwater photos of narwhals, although this is generally true of most arctic marine mammals and I for one don't blame photographers: it's damned cold! Photo by Paul Nicklen, National Geographic.

However, the above arguments follow for the narwhal: the tusks are indeed dimorphic, and if these functions are not important for females (85% of females lack erupted tusks, making sensory functions useless for nearly half of the species), they probably do not reflect the main purpose of the tusk. The extreme sexual dimorphism strongly indicates a social role, and another recent study (Kelley et al. 2014) has found a strong correlation between narwhal tusk size and testes mass – confirming the sexual/social importance of tusks. More observations of tusk use in the narwhal is necessary, but males have been observed rubbing or slapping tusks together, and broken tips of tusks have been found embedded in other male narwhal heads (and, heads of belugas) – indirect evidence of narwhal combat. Similarly, underwater observations of walrus and narwhal behavior and combat are rare or lacking altogether. 

Adorable bonus photo (by Paul Nicklen, Nationa Geographic/Getty Images).
What about other walruses?

Thus far, almost all discussions of tusk evolution in walruses have either been confined to the modern species, or daresay even cetaceans like Odobenocetops. Obviously, the former is a necessary starting point, and the latter merits consideration – but, what about extinct walruses? The only serious consideration of tusk evolution using fossil walruses was Deméré (1994), who (as outlined above) remarked upon tusks in walruses (e.g. Valenictus) from temperate and subtropical latitudes. An important question that hasn’t really been asked before is: who had the first tusks? The answer is remarkably easy and quick: the dusignathine Gomphotaria pugnax, which is 2-3 million years older than the earliest known tusked odobenine fossils. Tusks in Gomphotaria are quite a bit different in morphology than modern Odobenus: the tusks are short and procumbent, lack globular dentine, and a smaller pair of lower tusks are present; similar double-tusks are seen in Dusignathus (particularly D. seftoni). There is some variation even amongst the odobenines: Protodobenus has thickened maxillae and large canine roots, but the emergent canine crowns are barely proportionally larger than that in a sea lion; tusks are absent in Aivukus, and short, curved, and procumbent (forward inclined) tusks are present in Alachtherium/Ontocetus and Valenictus (although somewhat longer but no less precumbent). Morgan Churchill and I discussed a few of these points in our paper on Pelagiarctos (Boessenecker and Churchill, 2013). This pattern tells us several things: 1) “Sled runner” tusk function would have only really been present in the modern walrus, as most earlier forms had somewhat procumbent tusks that would not have been aligned with the seafloor; 2) tusks do not really seem to be correlated with any subset of the marine environment, and association with ice likely reflects a relatively recent (e.g. Pleistocene) adaptation of Odobenus to high latitude environments; and 3) tusks evolved in several directions in the last 8 million years, which if anything signifies sexual selection and recalls horn and antler diversity amongst small clades of sexually dimorphic and selective ungulates.

The moral of the story is this: there is a difference between what a structure evolved for and what its current function(s) is/are; when walrus tusks first evolved, there was no extensive pack ice and walruses inhabited temperate and subtropical latitudes. The walrus tusk continues to serve an important role in social behavior, but has been used for other purposes (locomotion, sleeping) and is thus an exaptation of sorts. This point can be extended to the narwhal: simply because the narwhal tusk can be sensitive to salinity and temperature does not mean that it evolved for that purpose. In both cases the evidence of sexual dental dimorphism is the most significant, and the evidence rather overhwhelmingly supports a social or sexual origin of tusks in both Arctic species.


R. W. Boessenecker and M. Churchill. 2013. A Reevaluation of the Morphology, Paleoecology, and Phylogenetic Relationships of the Enigmatic Walrus Pelagiarctos. PLoS One 8(1):e5411.

Deméré, T.A. 1994. Two new species of fossil walruses (Pinnipedia: Odobenidae) from the upper Pliocene San Diego Formation. Proceedings of the San Diego Society of Natural History 29:77-98

Geraads, D. 1997. Carnivores du Pliocene terminal de Ahl al Oughlam (Casablanca, Maroc). Géobios 30(1):127-164

Fay, F.H. 1982. Ecology and biology of the Pacific walrus Odobenus rosmarus divergens Illiger. North American Fauna 74:1-279.

Kelley, T.C., Stewart, R.E.A., Yurkowski, D.J., Ryan, A., and Ferguson, S.H. 2014. Mating ecology of beluga (Delphinapterus leucas) and narwhal (Monodon monoceros) as estimated by reproductive tract metrics. Marine Mammal Science (Online early: DOI: 10.1111/mms.12165

Miller, E.H. 1975. Walrus ethology 1. The social role of tusks and applications of multidimensional scaling. Canadian Journal of Zoology 53: 590-613.

Muizon, C. de. 1993. Walrus-like feeding adaptation in a new cetacean from the Pliocene of Peru. Nature 365-745-748.

Muizon, C. de., and Domning, D.P. 2002. The anatomy of Odobenocetops (Delphinoidea, Mammalia), the walrus-like dolphin from the Pliocene of Peru and its palaeobiological implications. Zoological Journal of the Linnean Society 134: 423-452.

Muizon, C. de., Domning, D.P., and Ketten, D. 2002. Odobenocetops peruvianus, the walrus-convergent delphinoid (Mammalia: Cetacea) from the early Pliocene of Peru. Smithsonian Contributions to Paleobiology 93: 223-261.

Nweeia, M.T., Eichmiller, F.C., Nutarak, C., Eidelman, N., Giuseppetti, A.A., Quinn, J., Mead, J.G., K’issuk, K., Hauschka, P.V., Tyler, E.M., Potter, C., Orr, J.R., Avike, R., Nielsen, P., and Angnatsiak, D. 2009. Considerations of anatomy, morphology, evolution, and function for the narwhal dentition. In Krupnik, I., Lang, M.A., and Miller, S.E. (editors), Smithsonian at the Poles: contributions to International Polar Year science. 223-240.

Nweeia, M.T., Eichmiller, F.C., Hauschka, P.V., Tyler, E., Mead, J.G., Potter, C.W., Angnatsiak, D.P., Richard, P.R., Orr, J.R., and Black, S.R. 2012. Vestigial tooth anatomy and tusk nomenclature for Monodon monoceros. The Anatomical Record 295:1006-1016.