Tuesday, August 26, 2014

The evolutionary history of walruses, part 2: the larger imagotariines: Pseudotaria, Pelagiarctos, Imagotaria, and Pontolis

For other entries in this series, see:

Walrus Evolution, part 1
Walrus Evolution, part 3
Walrus Evolution, part 4
Walrus Evolution, part 5

The previous post dealt primarily with the earliest diverging walruses – generally speaking, these were the smaller-bodied “imagotariines”. The “Imagotariinae” is a paraphyletic grade consisting of stem-walruses, and existed during the middle and late Miocene. The earliest diverging imagotariines detailed in the previous post include Prototaria, Proneotherium, Neotherium, and Kamtschatarctos. More derived imagotariines include Pelagiarctos from the middle Miocene of California, Pseudotaria muramotoi from the late Miocene of Japan, and Imagotaria downsi from the late Miocene of California and Oregon. Another problematic walrus is the giant Pontolis magnus from the late Miocene of Oregon, which may be an imagotariine, or possibly a dusignathine walrus – but will be discussed here rather than the next post on dusignathines.

Although small-bodied enaliarctine-like imagotariine walruses like Prototaria and Proneotherium are the earliest diverging walruses, the larger-bodied and later diverging walrus Pelagiarctos actually constitutes the earliest record of the family. Specifically, the mandibles of Pelagiarctos that Morgan and I published on in PLOS One (Boessenecker and Churchill, 2013) are from the “Topanga” Formation of Orange County, which is 17.5 Ma at the oldest, and thus either the same age as Prototaria from Japan and Proneotherium from Oregon, or slightly older. This suggests that when they first appear in the early Middle Miocene, there were already three genera of walruses and already a bit of size disparity.

The holotype skull of the imagotariine Pseudotaria muramotoi from the late Miocene of Hokkaido, Japan. From Kohno (2006).

Despite its smaller size and earlier diverging position on the cladogram, Pseudotaria muramotoi appears about 5-7 million years after Pelagiarctos, and was reported from the late Miocene of Hokkaido by Dr. Naoki Kohno (2006). Pseudotaria is intermediate in morphology and geochronologic age between the middle Miocene Neotherium and the late Miocene Imagotaria downsi. The holotype skull of Pseudotaria lacks the front of the rostrum and has no teeth, but is otherwise characterized by a wide and slightly arched palate, double-rooted cheek teeth, a braincase with a rounded lateral margin (e.g. lacking a box-shaped braincase of earlier odobenids and enaliarctines). Pseudotaria is more derived than Neotherium in the loss of a third root on the upper first molar, the loss of a furrow on the lateral side of the braincase corresponding to the pseudosylvian sulcus of the brain, and a slightly arched palate. However, more derived odobenids differ in having a wider and pentagonal basioccipital, and by lacking an upper second molar. Another imagotariine from the same formation representing a new genus is currently being described by my labmate Yoshi Tanaka and Dr. Naoki Kohno.

Specimens of Pelagiarctos sp. from the "Topanga Formation" (A) and Pelagiarctos from the Sharktooth Hill Bonebed (everything else).

Pelagiarctos is known only from a handful of specimens – the holotype of Pelagiarctos thomasi is a fragmentary “chin” consisting of left and right mandibles with poorly preserved teeth from the middle Miocene Sharktooth Hill Bonebed, collected by LACM head preparator Howell Thomas. Barnes (1988) also referred a handful of isolated cheek teeth to Pelagiarctos. Pelagiarctos thomasi is large – at least the same size as Imagotaria if not somewhat larger – and with noticeably larger cheek teeth than Imagotaria. Pelagiarctos thomasi is unique amongst stem walruses in having a fused mandibular symphysis; otherwise, the cheek teeth are double rooted and similar to Neotherium in shape by retaining a metaconid cusp (essentially, the teeth have four cusps), but are also inflated and bulbous, more similar to Imagotaria. Other features shared uniquely with Imagotaria amongst all walruses are the presence of a lingual cingulum (a ridge on the base of the crown along the lingual side – tongue-side – of the tooth) that bears numerous little cusps on it, and a canine that has a longitudinal groove on one side giving it a near figure-8 shaped cross section. Morgan Churchill and I (Boessenecker and Churchill 2013) referred a more completely preserved pair of mandibles with an incisor, canines, and three cheek teeth to Pelagiarctos – but owing to the unfused symphysis, we referred it to Pelagiarctos sp. This specimen also showed that Pelagiarctos primitively retains a second lower molar, like Proneotherium and Neotherium, and some specimens of Imagotaria, and unlike later diverging dusignathine and odobenine walruses. Because of the aforementioned similarities with Imagotaria, our cladistic analysis found that Pelagiarctos and Imagotaria form a clade – and depending upon future analyses, may permit redefinition of a more exclusive and monophyletic Imagotariinae (defined on at least a couple of dental features).

The "split" skull of Imagotaria downsi from the diatomite quarries of Santa Barbara County, referred by Mitchell (1968).

The sea lion-like walrus Imagotaria downsi was originally reported by Ed Mitchell in 1966 from late Miocene diatomite quarries in the vicinity of Lompoc in Santa Barbara County. The holotype specimen is from the latest Miocene Sisquoc Formation, while other nearby quarries yielding fossils of Imagotaria are from the underlying Monterey Formation. The stratigraphy of this area is problematic, as few quarries preserve the formational boundary between the Monterey and Sisquoc Formations, and both units are identical in lithology – meaning that age determinations need to be done using microfossils. Microfossil dates are only available for a couple of quarries – and microfossil dates for some of the more obscure quarries is no longer available, since some of these quarries are now closed and no longer exist.

An immature male skull of Imagotaria downsi from the Santa Margarita Sandstone near Santa Cruz, reported by Repenning and Tedford (1977).

The type specimen of Imagotaria is a partial rostrum and palate with much of an upper dentition, a nearly complete mandible, several isolated teeth, a fragmentary basicranium, atlas, humerus, and some other bits. Mitchell (1966) also referred a well-preserved complete skull (also from Lompoc) split horizontally and exposed in two slabs. I heard recently that during his Ph.D., Ed Mitchell wanted to reassemble the two slabs and prepare the specimen out and use that as the holotype – but then-director of the USNM, Remington Kellogg, would not allow the specimen to be reassembled. So, Mitchell chose the less complete specimen as the holotype. Additional specimens collected later from the Santa Margarita Sandstone in the Santa Cruz Mountains near Santa Cruz and the city of Scotts Valley from a series of sand and gravel quarries include two beautifully preserved skulls, mandibles, postcrania, and a couple of partial articulated forelimbs – all described by Repenning and Tedford (1977) and Barnes (1971). Imagotaria downsi is a relatively large pinniped, approximately the same size as an Australian sea lion (or, intermediate between a California sea lion and the Steller’s sea lion for fellow American readers). Like Pseudotaria, Imagotaria lacks the boxy braincase of earlier imagotariines, but differs from earlier imagotariines (Pseudotaria included) in having anterior premolars that are single rooted. Imagotaria differs from all earlier walruses in possessing a large and cuspate lingual cingulum (like Pelagiarctos) on the upper premolars in addition to having a well-developed protocone cusp on the cingulum. Imagotaria further differs from all earlier walruses in lacking the metaconid cusp on the lower cheek teeth. The available sample of Imagotaria indicates that there is a bit of variability in dental morphology – some specimens have a lower second molar and others do not, while some have two (or three) upper incisors, and there is some variation in postcanine tooth rooting; the available sample also shows evidence of sexual dimorphism. One specimen – a beautiful female skull from the Santa Margarita Sandstone reported by Repenning and Tedford (1977), known affectionately to some as “Rep’s girl”, has proportionally much smaller teeth than other specimens of Imagotaria downsi. A recent conversation with Dr. Naoki Kohno indicated that this specimen may represent a second species of small-toothed Imagotaria.

An adult female skull of Imagotaria from the same locality as the immature male, reported by Repenning and Tedford (1977) and affectionately known as "Rep's girl". This specimen may represent a second small-toothed species of Imagotaria.

Two partial articulated forelimbs (male, left, female, right) of Imagotaria downsi from the Santa Margarita Sandstone near Santa Cruz, including a tiny radius of Neotherium mirum for comparison (in middle); from Repenning and Tedford (1977).

Numerous additional but mostly undescribed specimens of Imagotaria have been reported from other late Miocene localities. Several skulls from the Empire Formation in Oregon represent a younger species of Imagotaria with single-rooted teeth – these specimens were briefly mentioned by Deméré (1994). I’ve seen them in person, and they are beautiful skulls, but much smaller than Imagotaria downsi. In the past few years, I’ve collected a few isolated teeth from the base of the Purisima Formation that are a dead ringer for Imagotaria – they have a cuspate lingual cingulum and a small pseudo-talonid basin, differing from the turnip-shaped teeth of dusignathine walruses (which also tend to lack enamel and have an inflated root). These teeth come from a bonebed that is just shy of 7 Ma, and represent the youngest fossil record of imagotariines anywhere.

The Lomita walrus. This thing is almost the same size as Pontolis and in life was likely just as scary. This fossil needs to be redescribed. From Lyon (1941).

One problematic Imagotaria-like walrus was reported and figured by Lyon (1941) from the late middle Miocene Valmonte Diatomite member of the Monterey Formation near Lomita in Los Angeles County, California. The Lomita walrus is huge - the skull is 56 centimeters long, and constitutes one of the largest described fossil pinnipeds in existence. Unfortunately, the skull is a bit squashed, but appears to have been relatively similar to Imagotaria and undescribed skulls of Pontolis magnus, but noticeably differs from both in exhibiting primarily single rooted teeth, and apparently lacking a cuspate lingual cingulum on the cheek teeth. Although this specimen was referred by Mitchell (1968) to Imagotaria downsi, this fossil is freaking enormous and about 20 centimeters longer than the largest known skulls of Imagotaria, in addition to lacking a cuspate lingual cingulum and a more derived pattern of root fusion. Instead, this monstrous walrus may be conspecific or congeneric with an undescribed and beautifully preserved skull and skeleton from the Capistrano Formation of Orange County known as “Waldo”, currently under study by student Isaac Magallanes at the Cooper Center (and recently highlighted in a recent talk by Isaac, Dr. Jim Parham, and myself at the Secondary Adaptations meeting in D.C.). Because study is still ongoing, I won’t spoil anything, other than just say “Wow” – you won’t be disappointed when it comes out, so stay tuned.

The holotype braincase of Pontolis magnus from the late Miocene Empire Formation of Oregon, photographed at the USNM in Washington D.C. That pinniped was pretty damn huge.

The last imagotariine, and by far the most impressive – is the gigantic walrus Pontolis magnus. Pontolis was described in 1905 by Frederick True, based upon a large braincase from the late Miocene (7-9 Ma) Empire Formation in Oregon. Since then, several specimens including complete and partial skulls, mandibles, and postcrania were referred to Pontolis by Deméré (1994) and one mandible was figured in our Pelagiarctos paper (Boessenecker and Churchill, 2013). Owing to ongoing work by Morgan Churchill and I, I won’t spoil many details other than what’s already been published. Pontolis is a giant imagotariine characterized by a huge nuchal crest on the braincase, large and ventrally flattened tympanic bullae, a long rostrum, incipiently single-rooted anterior cheek teeth, and an elongate mandible with a large digastric insertion and low, elongate coronoid process. Most impressive is the ridiculous size of Pontolis: its skull is 60 cm long, one-third larger than the largest modern walrus, and approaching the length of elephant seal skulls. Deméré (1994) originally found support for Pontolis as a dusignathine walrus, but a cladistic analysis by Kohno (2006) found Pontolis to form a clade with Imagotaria instead. Our cladistic analysis from the Pelagiarctos study found a compromise of sorts between the previous two studies: our results supported Pontolis as appearing one node closer to the “crown” Odobenidae – the Dusignathine + Odobeninae clade, and neither a dusignathine nor a sister taxon of Imagotaria. As always, more imagotariine (and dusignathine) fossils and a longer set of characters are needed to precisely pin down the phylogenetic position of Pontolis. I won’t say anything more, except that after two current otariid projects, Pontolis will likely be the next joint project by Morgan and I.

So, what does all this tell us? In general, late Miocene imagotariines had simpler dentitions, more robust skulls, larger canines (but not tusks, obviously), and more strongly concave (and more elongate) palates than the walruses of the middle Miocene. The late Miocene is notably the period in which the largest walruses existed, with two different forms approaching the size of elephant seals. Curiously, the largest imagotariines like Pontolis appear just as the similarly large desmatophocid Allodesmus became extinct – the youngest known record of Allodesmus is an undescribed skeleton from the late Miocene (9 Ma) Montesano Formation of Washington, USA. I have suspected since I was an undergraduate that imagotariines filled in available niche space after desmatophocids started on their way out.

Imagotariines of the late Miocene - look at that size disparity!  
Neotherium mirum for comparison (top left).

Rock units like the Empire Formation show us that at least two walruses – Imagotaria and Pontolis – were sympatric, and the only other described pinniped known from coeval rocks is the early fur seal Pithanotaria starri, an adorably tiny otariid about the size of a modern Guadalupe fur seal (Arctocephalus townsendi/phillippii). Altogether, this sort of diversity is reminiscent of pinniped diversity during the middle Miocene (e.g. the walruses Pelagiarctos and Neotherium, and the desmatophocid Allodesmus), but doesn’t compare well with the diversity of modern pinnipeds in the eastern North Pacific during the Holocene and late Pleistocene (four otariids and two phocids, as well as the walrus Odobenus straying far south in glacial periods). Notably absent during much of the late Miocene are tusked walruses – the first of which would appear about 7 Ma just before the close of the Miocene. The earliest Dusignathines also appear just before the end of the Miocene (unless Pontolis turns out to be a dusignathine after all). For whatever reason, imagotariines – which appear to reflect dietary generalists like modern sea lions – died out at the end of the Miocene, paving the way for strange tusked beasts with a taste for clams (and squid?). The last imagotariines, represented by a handful of teeth from the base of the Purisima Formation, mark the end of a ten million year dynasty of  generalist sea lion-like walruses – and mark the beginning of something a whole lot weirder.

Next up: the Dusignathines.


L. G. Barnes. 1971. Imagotaria (Mammalia: Otariidae) from the Late Miocene Santa Margarita Formation Near Santa Cruz California. PaleoBios 11:1-10.
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.

T. A. Deméré. 1994. The Family Odobenidae: A phylogenetic analysis of fossil and living taxa. Proceedings of the San Diego Society of Natural History 29:99-123.

N. Kohno. 2006. A new Miocene odobenid (Mammalia: Carnivora) from Hokkaido, Japan, and its implications for odobenid phylogeny. Journal of Vertebrate Paleontology 26(2):411-421.

G. M. Lyon. 1941. A Miocene sea lion from Lomita, California. University of California Publications in Zoology 47:23-41.

E. D. Mitchell. 1968. The Mio-Pliocene pinniped Imagotaria. Journal of the Fisheries Research Board of Canada 25(9):1843-1900.

C. A. Repenning and R. H. Tedford. 1977. Otarioid seals of the Neogene. Geological Survey Professional Paper 992:1-93.

F. W. True. 1905. Diagnosis of a new genus and species of fossil sea-lion from the Miocene of Oregon. Smithsonian Miscellaneous Collections 48(1):47-49.

No comments: