Saturday, March 21, 2009

The Ecomorphology of Arctodus

The enormous crown-group bear Arctodus simus is no stranger to this blog, I last covered it a couple years ago, but thanks to a new publication I feel obliged to give it another go. Other motivations include my guilt about letting that prior post get so bloated with topics and my annoyance with the portrayal of this animal in the media… so here we are!

Large and potentially fearsome animals are often subject to “Godzillafication” in popular culture and Arctodus simus is no exception. One of the more egregious offenders in recent memory is the oddly named show Jurassic Fight Club which claims that the bear is 11.5 feet tall (3.5 m), twice as big as extant bears at up to 2500 pounds (1134 kg), and capable of running at 45 miles per hour (72 km/h)! The “size charts” at that site are especially heinous and show a crudely rendered A. simus to have a shoulder height more typical for a proboscidean and its apparent nemesis Panthera atrox to be taller than most rhinoceroses. Illustrations by Mauricio Antón of both animals at a proper scale can be found here.


There is no denying that Arctodus simus was an enormous animal, but claims that it weighed a metric tonne or more are baseless. Christiansen (1999) used limb bone dimensions to calculate that a large male A. simus would weigh between 700 to 800 kg (1540-1760 lbs) and speculated that on rare occasions males could have reached or exceeded one tonne. I think visual comparisons can be highly valuable when dealing with speculative body masses and a mockup of a large A. simus and large polar bears (here) suggest that the estimates made by Christiansen (1999) are indeed rather excessive. Sorkin (2006) questioned the use of limb bone dimension regressions against mass since A. simus has proportionally long and robust legs; estimates based on condylobasal skull length gave a maxima of 570 kg (~1250 lbs) for the species. Since A. simus has a rather short skull I can't help but wonder how accurate condylobasal skull length would be in predicting mass - it should be noted that A. simus also has a short body and this weight estimate does at least look more on the mark visually.

Emslie and Czaplewski (1985) suggested that A. simus, which they though had a maxima of 620-660 kg, may have been beyond the physical limitations of an active predator. I'm not a big fan of using maxima, but since the maximal weights of unambiguous hypercarnivorous predators have been calculated they are actually useful in this situation. Panthera atrox has been given a maxima of 420 kg (925 lbs) based on condylobasal skull length and Smilodon populator "almost certainly exceeded" 400 kg (880 lbs) based on limb bone analysis (Sorkin 2008 - data from Turner and Antón 1997, Christiansen and Harris 2005). Since A. simus did not seem to be greatly in excess of those cats, I don't think active hypercarnivory can be ruled out on that basis.

A Herbivorous Relative

Apomorphies such as a premasseteric fossa and a skull with a short, deep, and convex shape place Arctodus simus in the New World bear subfamily Tremarctinae, which includes the extant spectacled bear Tremarctos ornatus (Trajano and Ferrarezzi 1994). Arctodus and the exclusively South American Arctotherium are united as a clade of "short-faced bears" on the basis of wide molars and presumably other traits (a more foreshortened face?) (Trajano and Ferrarezzi 1994, Soibelzon et al. 2005). Arctodus simus shares its genus with the poorly known A. pristinus, which appears to be distinguished primarily by being smaller in size and appearing slightly earlier in the fossil record (Dalquest and Mooser 1980, Puckette 1976). The Great Bear Almanac claims that other apomorphies of A. pristinus include a "slightly shortened" face, large and high-crowned teeth, and very large canines.

Emslie and Czaplewski (1985) interpret the premasseteric fossa, typically found in large herbivores and small carnivores, as related to an enlarged zygomaticomandibularis muscle which allows for lateral grinding and jaw elevation. A seldom mentioned character is the enlarged radial sesamoid of tremarctine bears, which is apparently a plesiomorphic character shared with giant pandas (Salesa et al. 2006). Its use in Tremarctos ornatus isn't too clear, but it could be related to extensive food manipulation and/or a rather arboreal lifestyle for a bear (Salesa et al. 2006). Emslie and Czaplewski (1985) suggest that the elongated limbs of A. simus could be used to pull down vegetation and I think the retention of an enlarged radial sesamoid could tie in nicely with this hypothesis. Whatever its function, an enlarged radial sesamoid does not sound like a likely characteristic for an alleged cursory animal...

skeleton from Emslie and Czaplewski (1985). Note the less extreme limb proportions and the enlarged radial sesamoid bone.

Lots of Unlikely Characters for a Hypercarnivore

Some early Arctodus simus workers interpreted the bear as having an ecomorphology more similar to a felid than an ursid on the basis of a short and broad rostrum, large carnassial blades, long limbs, and a "possibly" more digitigrade stance (Christiansen 1999 - citing Kurten 1967 and Kurten & Anderson 1980). This hypothesis is the one that got stuck in the popular imagination (wonder why...) and still is despite the fact that Sorkin (2006) utterly demolished it. The sectorial carnassials, position of the mandibular condyle, and development of the angular process which were previously interpreted as evidence for hypercarnivory are all present in Tremarctos, a near-exclusive herbivore (Sorkin 2006). Buccal cusps on the upper molars of A. simus are shared with the largely herbivorous Kodiak bear (Ursus arctos middendorfi) (Sorkin 2006). Additionally, Sorkin (2006) noted that relatively short canines, small orbits directed somewhat laterally, reduced leverage of deltoid and pectoralis muscles (used to subdue prey), reduced development of the brachialis and brachioradialis muscles (used to grasp prey), reduced pronation of the forearm and the flexation of the wrist and digits (used to grasp prey), a short and bent olecranon process (reduction of ability to crouch), brown bear-like lumbar vertebrae with limited flexion in the sagittal plane (meaning slower acceleration and top speed) all indicate that A. simus was not an active predator.

Matheus (1995) suggested that carbon and nitrogen isotope evidence indicated that A. simus was highly carnivorous, but functioned primarily as a scavenger adapted for long distance walking/predator intimidation. The only known vertebrate obligate scavengers are large soaring fliers and while a large terrestrial scavenger is not energetically infeasible per se, the hypothetical niche could only have evolved in an environment without vultures (Ruxton and Houston 2004). Furthermore, the isotope range overlaps with Pleistocene brown bears (Ursus arctos) and it does not appear that a hypercarnivore and omnivore with a diet including terrestrial mammals can be distinguished from isotopes (Sorkin 2006).

Warp Analysis

One of the aforementioned inspirations for this post was another recent re-evaluation of A. simus morphology by Figueirido et al. (2009), which I can't believe I'm just getting around to now. Modern faunivorous bears, the insectivorous sloth bear Melursus ursinus and the carnivorous polar bear Ursus maritimus, have long and flat skulls with frontalized orbits, downwards directed zygomatic arches, a slender jaw with small movement arms for the temporalis and masseter muscles and poorly developed upper and lower tooth rows (Figueirido et al. 2009). The vastly different prey of the two bears and their apparent convergences would seem to suggest that if A. simus were hypercarnivorous, it would at least display some of these characters. Herbivorous bears, the spectacled bear and the giant panda, have a short jaw with large movement arms for the temporalis and masseter, a horizontal ramus deep at the level at the level of the third molar and shallow below the canines, well-developed cheek teeth, small canines, well-developed zygomatic arches, and lateralized orbits (Figueirido et al. 2009). Even though those are the two most basal bear species, the fossil record appears to indicate that they developed these traits independently rather than retained them (Figueirido et al. 2009). Arctodus simus does not group with either of these ecomorphs, but falls in the middle with generalized omnivorous ursid species (Figueirido et al. 2009). Sorkin (2006) reached a similar conclusion and thought the bear could be compared with the striped and brown hyenas (Hyena hyena and Hyena brunnea) and may have had a diet consisting of large animal carrion, small animal prey, and plant material. Figueirido et al. (2009) suggest an even more thorough ecomorphological study including other carnivorans should be done with A. simus, but conclude that it was likely an omnivore capable of dealing with with seasons and climactic cycles. Of course, even animals with a highly plastic diet aren't immune to extinction.

So that's about all I want to say about Arctodus simus - I have no idea why this post took to write as long as it did. There is little doubt that some less-enlightened members of the media will continue to erroneously portray this animal as a hypercarnivorous terror, but hopefully the alternate (and more parsimonious) suggestions will gain some attention.


Carbone, Chris et al. 2007. The Costs of Carnivory. PLoS Biology. Available

Christiansen, Per and Harris, John M. 2005. Body size of Smilodon (Mammalia: Felidae). Journal of Morphology 266, 369-384

Christiansen, Per. 1999. What size were Arctodus simus and Ursus spelaeus (Carnivora; Ursidae)? Ann. Zool. Fennici 36, 93-102. Available

Dalquest, Walter W. and Mooser, O. 1980. Arctodus pristinus Leidy in the Pleistocene of Aguascalientes, Mexico. Journal of Mammalogy 61, 724-725.

Emslie, Steven D. and Czaplewski, Nicholas J. 1985. A new record of the giant short-faced bear, Arctodus simus, from western north America with a reevaluation of its paleobiology. Natural history museum of Los Angeles county contributions in science 371, 1-12.

Figueirido, B. et al. 2009. Ecomorphological correlates of craniodental variation in bears and paleobiological implications for extinct taxa: an approach based on geometric morphometrics. Journal of Zoology 277, 70-80.

Matheus, Paul E. Diet and Co-Ecology of Pleistocene Short-Faced Bears and Brown Bears in
Eastern Beringia. Quaternary Research 44, 447-453

Puckette, William L. 1976. Notes on the Occurrence of the short-face bear (Arctodus) in Oklahoma. Proc. Okla. Acad. Sci. 56, 67-68

Ruxton, Graeme D. and Houston, David C. 2004. Obligate vertebrate scavengers must be large soaring fliers. Journal of Theoretical Biology 228, 431-436

Salesa, M. J. et al. 2006. Anatomy of the “false thumb” of Tremarctos ornatus (Carnivora, Ursidae, Tremarctinae): phylogenetic and functional implications. Estudios Geológicos 62, 389-394

Soibelzon, Leopoldo, H. et al. 2005. The fossil record of South American short-faced bears (Ursidae, Tremarctinae). Journal of South American Earth Sciences 20, 105-113.

Sorkin, Boris. 2008. A biomechanical constraint on body mass in terrestrial mammalian
predators. Lethaia 41, 333–347

Sorkin, Boris. 2006. Ecomorphology of the giant short-faced bears Agriotherium and Arctodus. Historical Biology 18, 1-20

Trajano, E. and Ferrarezzi, H. 1994. A fossil bear from Northeastern Brazil, with a phylogenetic analysis of the South American extinct Tremarctinae (Ursidae) Journal of Vertebrate Paleontology 14, 552-556

Wednesday, March 4, 2009

Giant Leptocephali

Elopomorpha is a diverse group of basal teleost orders united by a leaf-like larval form known as a leptocephalus; members include the tarpons and relatives (Elopiformes), bonefish (Albuliformes), halosaurs and spiny eels* (Notacanthiformes), true eels (Anguilliformes), and the gulper eels and relatives (Saccopharyngiformes) (Inoue et al. 2004). Inoue et al. (2004) finally corroborated the monophyly of Elopomorpha with character matrices and determined this phylogeny:

*Halosaurs and spiny eels have been placed in o
rder Albuliformes by Inoue et al. (2004).

Fig. 5 from Inoue et al. (2004). Elops and Megalops are Elopiformes, Albula and Pterothrissus are members of Albulidae (bonefish), Aldrovandia and Notacanthus are members of Halosauridae and Notocanthidae, Gymnothorax through Anguilla are Anguilliformes, Saccopharynx and Eurpharynx are Saccopharyngiformes. Note that Albulidae and Anguilliformes are paraphyletic according to this phylogeny. The leptocephalus tail shape is important later on in this post.

Leptocephali are not normally very large, typically some 2-4 inches (50-100 mm) in length, so specimens substantially larger than this certainly raise eyebrows. The most commonly mentioned giant leptocephali were those collected by the Dana, the largest of which measured around 1800 mm (Castle 1959, citing Bertin 1954). Unfortunately, I have not found any information on these specimens aside from the length of the one individual and the number of myomeres (450*); their current whereabouts appear to be unknown. The ambiguity of this report has not stopped some cryptozoologists (e.g. Heuvelmans) from assuming that the larvae were true eels** with Anguilla-like growth (45 mm leptocephalus to a 800 mm adult) and a ~30 m adult size! Castle (1959) pointed out that Nemichthys scolopaceus has a larvae up to 253 mm in length and a growth rate which would turn a 1.8 m larvae into a 2.7-5.4 m adult. As we'll see, it's not entirely out of the question for the 1.8 m larvae to have an adult length shorter than that of the larvae.

*The number of myomeres apparently led Bertin to think that the larvae were members of Nemichthys (Castle 1959).

** Castle (1959) seems to synonymize "leptocephalid" with "eel larvae", so apparently he was unaware of the larval stage occuring in several other orders. Elopomorpha was established in 1966 on the basis of leptocephalus larvae (Inoue et al. 2004), and it seems improbable that four leptocephalus-adult relationships were worked out in 7 years. Heuvelmans, writing in 1968, certainly should have known.

Leptocephalus giganteus was first described by Castle (1959) on the basis of a 893 mm leptocephalus caught in shallow water off South Westland, New Zealand. The larvae was extremely attenuated with an 18.1 mm head, pre-vent length of 883 mm, depth of 18 mm, and 466-486 myomeres (Castle 1959). Castle (1959) suggests that characteristics such as the lack of fin rays, lack of cranial and pectoral development, V-shaped myomeres, and a subterminal vent indicate that this larvae is not a pathological giant and has not yet reached its full growth. He concluded that there's a giant bathypelagic eel living off the coast of New Zealand...

Another L. giganteus from S. Africa was described in 1967 and two more were reportedly found in an Alepisaurus* stomach, but it wasn't until Smith (1970) that an incredibly important feature was recognized. True eels never have pelvic fins at any of their life stages and sure enough, Smith (1970) noticed barely visible pelvic fins at the 34th myomere of his large individual. Coupled with the short-based dorsal fin, L. giganteus was recognized as a notacanthiform (Smith 1970). Interestingly, this excerpt from Early Stages of Fishes in the Western North Atlantic puts the maximum length of L. giganteus at 1840 mm (or 2 m), awfully close to the reported length of the largest 1930 Dana specimen. Are there reasons for thinking that the Dana specimens are L. giganteus aside from the length and myomere count (300-486 in L. giganteus, 450 in Dana)? Even more curiously, the book cites a source which suggests that L. giganteus may be a synonym for Notacanthus chemnitzi; are size and a widespread distribution the reasons for thinking this? I found a Moser and Charter (1996), but it did not have any unique information on L. giganteus...

*What's with all these fish having a suffix that means "lizard"?

Giant leptocephali appear to have been known since 1810 (from Rafinesque) and these unusual elongated larvae were eventually placed in the "genera"/categories Tilurus and "Tiluropsis"* by Roule (Smith 1970). While L. giganteus has an elongated head and round eyes, Tilurus can be distinguished by its short head and round eyes and "Tiluropsis" has a short head with vertically elongated eyes (Smith 1970). "Tiluropsis"-type larvae were recently captured in Brazilian waters, giving the authors an opportunity to discuss how prior workers tenuously connected the larvae to the halosaur Aldrovandia gracilis; the two have similar branchiostegal and pectoral ray counts and overlapping distributions, but the authors feel that much work is needed to clear up the life cycle of the enigmatic larvae. The book excerpt suggests that Tilurus belongs to the other notacanthiform subgroup, Notacanthidae.

* Tilurus has nomenclatural standing and "Tiluropsis" doesn't - somehow. Another type, "Tilurella" turned out to be a post-larval Nemichthys - it was distinguished by having elongated dorsal and anal fins, an anguilliform nasal organ, et cetera (Smith 1970).

With the advent of molecular data, if someone manages to get hold of giant leptocephalus and notacanthiform tissues they could theoretically resolve the mysteries surrounding the larvae. Do the larvae shrink in size at transformation? How many species compose each of the leptocephalus types? Do all notacanthiform larvae have this elongated shape? Although I'm somewhat doubtful that any of these will turn out to be a 30 m eel, it should still be a fascinating study nonetheless...


Castle, P. H. J. 1959. A Large Leptocephalid (Teleosti, Apodes) from off South Westland, New Zealand. Transactions of the Royal Society Society of New Zealand 87, 179-184. Available

Figueroa, Daniel E. et al. 2007. The southernmost record of notacanthiform Tiluropsis leptocephali, with notes on possible species identity. JMBA2 - Biodiversity Records. Available

Inoue, Jun G. et al. 2004. Mitogenomic evidence for the monophyly of elopomorph fishes (Teleostei) and the evolutionary origin of the leptocephalus larvae. Molecular Phylogenetics and Evolution 32, 274-286.

Smith, David G. 1970. Notacanthiform Leptocephali in the Western North Atlantic. Copeia 1970, 1-9.