Tuesday, February 12, 2013

Was Pelagiarctos a "killer" walrus? Part 5: life restoration

Now that I’ve published research that made the front page of the fox news website, I can consider myself satisfied as a scientist. The Pelagiarctos article has made some additional press, including Science World Report, the Orange County Register, the Otago Daily Times (our local newspaper), and my favorite popular article, titled "Ancient "killer walrus" cuter than originally thought." Also, take a look at Brian Switek's write up on Laelaps.

It's in there, in between the articles about the second amendment and how the government 
will be coming after you.

I thought I’d go ahead and explain in a post about the making of the life restoration of Pelagiarctos, and discuss what it may have looked like in life. As detailed in our phylogenetic analysis, Pelagiarctos is most closely related to the late Miocene walrus Imagotaria downsi. Imagotaria downsi is known from the Santa Margarita Sandstone and Sisquoc Formation of Northern and southern California (respectively), and is early late Miocene in age (Tortonian stage – 9-12 Ma). The type specimen of Imagotaria downsi is sort of a cruddy specimen, but a beautiful collection of well preserved fossils was reported by Charles Repenning and Richard Tedford in 1977 from the Santa Margarita Sandstone near Santa Cruz, California. Two skulls including an adult female (“Rep’s Girl” as some marine mammal paleontologists call it) and a subadult male are sea lion like in their morphology. They have long snouts, the skulls are flat-topped, and have low saggittal crests, large canines, and deep mandibles. Although superficially sea lion-like in general form, they lack the supraorbital shelves typical of otariids.

The female skull of Imagotaria downsi from Santa Cruz, affectionately known as "Rep's girl".

The lack of similarity between Pelagiarctos and the modern walrus is evident in its morphology. In the modern walrus, the canines are reduced in size to small pegs, no larger than the premolars (and are thus called ‘premolariform’); the incisors are totally absent, and the “chin” of the modern walrus mandible tapers to a triangular point that lacks teeth and instead has a longitudinal furrow. The transverse tapering of the mandible accommodes the hugely expanded tusks, which for the uninitiated, are the upper canines. The reduction in size of the lower canines presumably also permits enlarged upper tusks – while the complete loss of incisors probably allows an unobstructed pathway for suction into the oral cavity (see the section on odobenid dental and mandibular evolution in Odobenidae in our paper).

Rough line drawing of "Rep's girl".

Instead, Pelagiarctos – like Imagotaria – appears to be rather sea lion-like in overall morphology, perhaps something similar to one of the larger, more robust sea lions such as the New Zealand sea lion (Phocarctos hookeri), South American sea lion (Otaria byronia), or Steller’s sea lion (Eumetopias jubatus); the mandible of California sea lions is noticeably smaller and less ‘deep’ than Pelagiarctos. So, we have an overall idea of the shape of the head of Pelagiarctos. To start, I took a line drawing of the skull and mandible of Imagotaria downsi, and in adobe illustrator reduced them to the same mandible size. Then, I took the facial part of the skull, and shortened it to fit the short toothrow of Pelagiarctos. Figuring that the rostrum of Pelagiarctos would have probably been deeper and more robust like the mandible, I also made the facial region more dorsoventrally deep.

Cranial reconstruction of Pelagiarctos sp. based on the proportions of Imagotaria downsi.

Then, using this new skull reconstruction, I sketched it in an oblique view by using a reference photo of an Imagotaria skull in the same oblique view, while adding in the changes in proportions. I made sure to sketch it in an angle where I had a photograph of the new Pelagiarctos specimen, so that I could later on digitally overlay the photo of the mandible.

What did Pelagiarctos look like? More like a sea lion, or a fellow walrus? Due to its early position and early timing in pinniped evolution, it probably had some sort of external ears. It was not a gigantic cold-adapted pinniped like the walrus, so it may have primitively retained abundant fur or hair. It's long whiskers as I've reconstructed it are consistent with a piscivorous habit, rather than the molluskivorous feeding behavior of the modern walrus. My initial black and white graphite drawing is shown here.

So, we’ve got a general shape of the skull and head – but what would it look like? There are still a couple more considerations. For example – would it have had fur? Long vibrissae (whiskers), or short vibrissae? Would it have had thick blubber instead, like the modern walrus? And what about ears – modern walruses (and true seals) don’t have external ears, but sea lions and fur seals (Otariidae) have dinky little ear flaps. Before I continue with this discussion, I must stress that this is all highly speculative. Using the “extant phylogenetic bracket”, and assuming that molecular analyses have correctly identified sea lions and walruses as sister taxa, we can infer that it would look closest to a sea lion or a modern walrus. Okay, that basically includes all of the aforementioned features. The modern walrus is technically closer – but it is a highly derived animal, whereas Pelagiarctos is a very generalized sea lion-like pinniped. I reconstructed Pelagiarctos with external ear flaps to reflect the fact that most early pinnipeds probably had ear flaps or even large external ears (e.g. like an otter). After all, external ears are primitive, and it would be silly to assume that true seals and walruses have lacked ears throughout their evolutionary history. Considering the molecular support for a sea lion + walrus clade, it appears that external ear loss is convergent in the walrus and true seals anyway. What about fur, then? Only a few pinnipeds truly lack dense fur or hair – the walrus, and the elephant seals. The southern elephant seal and the walrus are both high latitude, cold water adapted - but they are also substantially larger than Pelagiarctos. Given the temperate latitude and similarity in size of Pelagiarctos with modern sea lions, which lack fur but have dense hair – it can be parsimoniously reconstructed as “fuzzy”. On that note, I really ought to talk about Heather Liwanag's awesome study on the evolution of marine carnivore fur/hair. Lastly, I reconstructed it with long whiskers because it’s a pelagic hunter – the short, stubby whiskers of the modern walrus are an adaptation for “feeling” benthic invertebrates and sediment.

And the final reconstruction, all colorized and everything. I had a lot of fun doing this reconstruction, and it seemed to do the trick.

Anyway, this concludes my series of posts on the new study by Morgan and I. Hope you enjoyed reading it (and hopefully the actual paper as well).


Boessenecker, R.W. and M. Churchill. 2013. A reevaluation of the morphology, paleoecology, and phylogenetic relationships of the enigmatic walrus Pelagiarctos. PLoS One 8(1) e54311. doi:10.1371/journal.pone.0054311.

Sunday, February 10, 2013

Epic fossil whale collage

I walked into my labmate Cheng-Hsiu Tsai's office recently and saw that he had a new desktop background - there were some specimens I didn't even recognize. It's a collage he made with nearly every published fossil mysticete he could find (and some modern mysticetes). Tsai, who is currently on a two month research trip to Taiwan, Japan, and Australia - said it would be OK for me to share it here with you all - so, enjoy!

Thursday, February 7, 2013

Was Pelagiarctos a "killer" walrus? Part 4: No, probably not. The feeding ecology of Pelagiarctos reassessed

Sorry for the short break – I’ve been fairly busy the last two weeks doing a number of things, but most of all revising a manuscript that was accepted last week on barnacle-encrusted sea lion bones from Oregon. I’m also pleasantly satisfied with additional press attention our Pelagiarctos article got over the last couple of weeks, and the PLOS metrics indicates it’s already gotten 3,000 views, which is quite a few more than my 2011 fur seal article (under 100), although that’s probably because it’s 1) linked to in all the news articles and 2) not paywalled.

Comparison of the holotype (B, C) and referred mandible (A) of Pelagiarctos. From Boessenecker and Churchill (2013).

When Morgan and I started our research on the new specimen of Pelagiarctos, we realized that the most interesting application of the research would be reevaluating the interesting hypothesis by Barnes (1988) that Pelagiarctos thomasi was a specialized macrophagous predator. To recap, Barnes interpreted this novel hypothesis based on several lines of evidence: 1) Pelagiarctos thomasi is relatively rare in the Sharktooth Hill Bonebed relative to other pinnipeds (apex predators are rare because they require a large population of prey items to subsist upon), 2) sharp postcanine teeth roughly similar in morphology to those of bone-cracking hyaenids and borophagine canids, 3) a fused mandibular symphysis, suggesting an adaptation towards high bite force, and 4) large body size. In addition to these interpretations, Barnes (1988) further interpreted all the isolated teeth as being from males – as the holotype “chin” has large canines, and at least one of the postcanine teeth slipped right in to the empty alveolus of the holotype. The rest of the postcanines are of similar size, suggesting to Barnes that they were all from one gender. Most pinnipeds – and all fossil walruses for which we have sufficient sample sizes (even wee little Proneotherium) are sexually dimorphic, with larger males. Furthermore, one of the holotype canines was broken in life and then worn from continued use, which Barnes (1988) interpreted as the result of male-male combat. Barnes further speculated that the lack of females could be caused by geographic separation of sexes – certainly an intriguing possibility, but difficult to test with such a small sample size.

An example of a vagrant pinniped: a leopard seal that wound up on a New Zealand beach. Apparently this happens somewhat often, but regardless - New Zealand is not in the normal range of this animal. From keaphotography.org

We identified several other hypotheses which could just as parsimoniously explain the rarity of Pelagiarctos within the Sharktooth Hill Bonebed. Given the extremely large sample size of fossil vertebrates from the bonebed, and a century of intensive collecting, the rarity is probably a real phenomenon and thus probably not a result of preservational bias (least of which because Pelagiarctos is a large animal and has a higher preservation potential than the smaller but numerically more abundant Neotherium). Numerous studies of modern pinnipeds have demonstrated that they are prone to vagrancy – in the ocean, after all, it’s easy to get caught up in currents or forage further away than other members of your species. It’s also possible that Pelagiarctos was simply a pelagic, offshore pinniped, rarely straying into coastal waters off Orange County (“Topanga” Formation) or the Temblor Sea (Sharktooth Hill). Furthermore, the Sharktooth Hill Bonebed was deposited over a protracted period of time (~700,000 years) due to a depositional hiatus, and it’s possible that a short period of time could have seen introduction of Pelagiarctos (from further south, north, or further out in the Pacific) along with a brief change in climate, ecology, or circulation.

A sea lion and a fishy smorgasbord. Photo by David Doubilet.

The fused mandibular symphysis is a bit more ambiguous. The only modern pinniped with a fused symphysis is the extant walrus Odobenus rosmarus, and it’s not immediately clear why, or how it could be adaptive relative to feeding. My hunch is that it’s a secondary consequence of having a pachyosteosclerotic mandible – the lower jaw of Odobenus is thickened around the chin and dense, and development of this may have resulted in symphyseal fusion. The earlier walrus Alachtherium/Ontocetus has an unthickened chin and lacks fusion – but Valenictus chulavistensis (the strange sister taxon of Odobenus) has a fused symphysis, and also lacks a thickened “chin”. Extreme pachyosteosclerosis of the skull and mandible in Odobenus has been suggested as a possible adaptation for keeping the head negatively buoyant during benthic foraging. One other fossil odobenid has a fused symphysis – Dusignathus seftoni from the San Diego Formation. It’s not even clear what Dusignathus ate, so it’s not a very good analogue either. Barnes (1988) argued that a fused symphysis in Pelagiarctos thomasi suggested and adaptation for large bite forces – however, carnivorans with high bite forces such as borophagine canids and hyaenids (which Barnes compared the dentition of Pelagiarctos to), as well as sea otters – all have unfused symphyses, possibly to allow slight movement of the mandibles so as to avoid tooth damage (Scapino, 1981). So what was mandibular fusion in Pelagiarctos thomasi for? Who knows! That’s for someone else to figure out. Besides, the new specimen didn’t have a fused symphysis anyway.

Comparison of calculated trophic level and body mass in pinnipeds. Lower trophic level corresponds to feeding upon benthic invertebrates, and high trophic level corresponds to eating large fish and cephalopods. From Boessenecker and Churchill (2013).

A walrus after a successful seal kill. These events have only been witnessed a few times by humans. Apparently, walruses will use their powerful suction normally reserved for mollusks to literally suck the meat right from the bones. From moblog.net.


 Another gory shot of a walrus feeding on a poor seal. From Vlasman and Campbell, Diseases and Parasites of Mammals of the Eastern Arctic.

We reinvestigated the issue of body size as well. Morgan has been working on a method to estimate the body mass of fossil pinnipeds, and his preliminary results indicated that the length of the lower toothrow is the single best predictor of body mass in a dataset of modern pinnipeds (fortunate for us, since all we had was a mandible to work with). Morgan was able to estimate the body mass of Pelagiarctos sp. at approximately 350 kg (~770 lbs), which is similar to some modern sea lions (male South American sea lions, California sea lions). There are much larger sea lions, however – the Steller’s sea lion (Eumetopias jubatus), which weigh up to 1,150 kg (2,500 lbs). Steller’s sea lion is the fourth largest pinniped (after the two elephant seals and the modern walrus), but is not a macrophagous predator: it feeds predominantly on fish, although it will occasionally prey upon juvenile pinnipeds. The modern walrus also occasionally preys upon juvenile seals and marine birds – yet it is clearly adapted and specialized for mollusk predation. All large bodied otariids predominantly eat fish as well, and many adult male sea lions of other species as well will also occasionally consume warm blooded prey when given the opportunity. Morgan further investigated this by calculating the trophic level of modern pinnipeds and plotting it relative to body mass. The result is that there is no apparent trend between body mass and diet – with one exception: the largest pinnipeds fed both at the high and low trophic levels (e.g. fish, cephalopods, as well as benthic invertebrates – e.g. walruses and bearded seals). In other words, large body mass doesn’t necessarily indicate anything specific about feeding ecology or diet. A previous analysis by Peter Adam and Annalisa Berta (2002) only found a very poor correlation between morphology and diet. That evidence of adaptations for macrophagy is lacking within pinnipeds is highlighted by the leopard seal: it does not use its postcanine teeth to feed upon penguins and seals, and only uses them for filter feeding for krill the rest of the year (8 months or so out of the year). When it feeds upon large bodied prey, it nips with its incisors and canines – which are not really any different from those of fish-eating pinnipeds. In other words, the only modern pinniped which could be argued to be a macrophagous apex predator only has dental specializations for feeding upon krill.

Arguably the only macrophagous pinniped - the leopard seal spends most of the year eating krill, and doesn't use its delicate postcanine teeth for killing penguins and seals, and instead only for filter feeding. Figures from Hocking et al. (2013).

So, not a "killer" walrus after all, but still a pretty intimidating beast.

So what did Pelagiarctos feed on? Probably fish, cephalopods, the normal menu for large bodied pinnipeds. It very well probably did feed upon warm blooded prey – occasionally, anyway (again, like modern pinnipeds). We’re not arguing that Pelagiarctos did NOT eat warm blooded prey – rather, we’ve made the case that Pelagiarctos lacked any adaptations which would lend themselves to macrophagy. One last point of interest – within the Sharktooth Hill Bonebed, Pelagiarctos is not even the largest pinniped; it’s dwarfed by Allodesmus, which Morgan estimated at 1400 kg (~3,000 lbs)! That’s just enormous (and approximately ¾ the size of modern northern elephant seals). Allodesmus, on the other hand, had enormous orbits possibly indicating deep diving adaptations (to which its large body size may have helped with as well), and a long snout with simple teeth – a definite contrast to the short “bulldog” face and dentition of Pelagiarctos.


Adam PJ, Berta A (2002) Evolution of prey capture strategies and diet in the Pinnipedimorpha (Mammalia, Carnivora). Oryctos 4: 83–107.

Barnes LG (1988) A new fossil pinniped (Mammalia: Otariidae) from the middle Miocene Sharktooth Hill Bonebed, California. Contributions in Science, Natural History Museum of Los Angeles County 396: 1–11.

Boessenecker RW, Churchill M (2013) A Reevaluation of the Morphology, Paleoecology, and Phylogenetic Relationships of the Enigmatic Walrus Pelagiarctos. PLoS ONE 8(1): e54311. doi:10.1371/journal.pone.0054311

Hocking, D.P., Evans, A. R., and E.M.G. Fitzgerald. 2013. Leopard seals (Hydrurga leptonyx) use suction and filter feeding when hunting prey underwater. Polar Biology 36:2:211-222.

Scapino R (1981) Morphological investigations into functions of the jaw symphysis in carnivorans. Journal of Morphology 167: 339–375