Just some of South America's Cenozoic megafauna (more complex versions of this illustration will appear later, and a larger version at different format is shown down below). These animals were not all contemporaneous. L to r: teratornithid, phorusrhacid, thylacosmilid borhyaenoid, toxodontid toxodont, astrapothere (at back), isotemnid toxodont, prothylacinid borhyaenoid, sebecosuchian.


I’ve often (or sometimes) said that there are – still, even after more than six years of operation – whole groups of tetrapods where I’ve barely scratched the surface, if that. The recent revisiting of borhyaenoids reminded me how much I love South American Cenozoic megafauna, and how frustrating it is that data on these animals can only be found in the technical literature and is almost never synthesised or made in the least bit popular.

Available on amazon (which is where this cover image comes from).

I should add that this isn’t because the respective specialists and aficionados are uninterested in such an aim – it’s because synthesising the scattered and obscure literature on these animals is difficult, and there’s not all that much incentive to do it (publishers and the public at large display little interest in these animals). The best synthetic treatment remains G. G. Simpson’s 1980 Splendid Isolation, the Curious History of South American Mammals (Simpson 1980), and if you’re seriously interested you absolutely must get hold of this book. Savage & Long’s (1986) Mammal Evolution: An Illustrated Guide isn’t too bad on South American native mammals, and the recent multi-authored volume Bones, Clones & Biomes: The History and Geography of Recent Neotropical Mammals (Patterson & Costa 2012) includes a lot of relevant data and discussion.

As you’ll recall from the borhyaenoid article, island South America was inhabited by diverse metatherian predators. These lived alongside the flightless phorusrhacid birds, an array of xenarthrans (sloths, armadillos, glyptodonts and kin), and an enigmatic set of mostly herbivorous placental mammals – the pyrotheres, astropotheres, xenungulates, didolodontids, litopterns and notoungulates.

What, exactly, is a notoungulate anyway?

Exactly where these placental mammal groups fit within the mammalian family tree remains uncertain, though there are several ideas. In this article I want to start a long-overdue and much anticipated look at notoungulates. There’s so much to say about this group that I can only deal with one small set of species at a time, and here I want to discuss members of just one notoungulate group, Toxodontia. And - because there’s an awful lot to say about toxodonts in general - this article is devoted to one toxodont group alone, Isotemnidae. We’ll get to the others later. How, I hear you ask, are notoungulates identified, and how are toxodonts identified?

Notoungulate upper right molar showing key structures. After diagram in Croft (1999).

The enormous number of taxa grouped together in Notoungulata share an unusual set of cranial, dental and ankle characters, now regarded as synapomorphies that support the recognition of this group as a clade. In the notoungulate skull, the upper molars have what’s called a ‘coronal pattern’: the cusps have been transformed into long crests arranged around the edges of the tooth (the crests are called lophs on the upper teeth, and lophids on the lower ones). A protoloph borders the anterior edge of the tooth, a metaloph along the posterior one, and an ectoloph runs parallel to the tooth’s labial (outer) side. A unique accessory spur – the crochet – projects from the metaloph in the direction of the ectoloph (Simpson 1980, Cifelli 1993, Croft 1999). As usual with tooth cusp characters, this pattern was extensively modified in advanced lineages.

Several key features concern the ear region and the bones surrounding it. The zygomatic arch in the notoungulate skull is attached very high up on the side of the skull [see Toxodon skull shown below], an ‘epitympanic sinus’ is present in the squamosal bone at the back of the skull, a large, bony chamber (an auditory bulla) projects downwards beneath the ear, and the auditory meatus (the bony ear opening itself) is tubular and associated with a bony crest (Cifelli 1993, Gabbert 2004). As usual with weird anatomical characters, this set of features surely means something very interesting but (so far as I know), nobody has ever worked out exactly what that is. Sensitive, low-frequency (infrasonic?) hearing has been suggested (Savage & Long 1986).

Skull of the toxodontid toxodont Toxodon, photographed at Zoologisk Museum, Copenhagen, by FunkMonk. Total length of skull = c. 60 cm. Licensed under Creative Commons Attribution-Share Alike 3.0 Unported license.

Within notoungulates, a lengthened molar series, a shelf-like process on the lower premolars (technically, an anterolabial cingulum) and the presence of a lower molar crest called a paralophid unite the toxodonts. Toxodonts are best known for the gigantic (hippo- or rhino-sized) toxodontids of the Miocene, Pliocene, Pleistocene and Holocene. But this is only one of several groups, and in this article we’re going to start going through them. I should note that a number of toxodont taxa – mostly (but not entirely) from the Paleocene and Eocene – have been hypothesised to be ‘basal toxodonts’, outside the clade containing all other species (McKenna & Bell 1997). They include Brandmayria, Colhuelia, Lafkenia, Colhuapia, Allalmeia, Brachystephanus and Xenostephanus.

Introducing the isotemnids

Reconstruction of Thomashuxleya (though see text!), from Simpson (1967).

About 12 toxodont genera from the Paleocene, Eocene and Oligocene are grouped together in Isotemnidae. None can be considered familiar, though a near-complete skeletal reconstruction produced for the Casamayoran* Patagonian form Thomashuxleya externa by Simpson (1967) - shown above - has formed the basis for illustrations in various popular books. It is in fact one of the most familiar toxodont reconstructions in the literature, perhaps being second only to Owen’s depiction of the gigantic toxodontid Toxodon.

* Casamayoran = middle or Late Eocene. One of about 21 terms used in the SALMA (= South American Land Mammal Ages) biochronology.

Thomas H. Huxley photographed c. 1870. From wikipedia, in public domain.

The name Thomashuxleya of course honours Victorian biologist and ‘Darwin’s bulldog’ Thomas H. Huxley. Florentino Ameghino (1854-1911), the great Argentine palaeontologist, naturalist, zoologist and anthropologist who described and named so many South American fossil mammals, was clearly fond of naming taxa after palaeontological colleagues: he also came up with Oldfieldthomasia, Ricardolydekkeria, Guilielmofloweria, Asmithwoodwardia, Carolodarwinia, Ricardowenia and others. Note that all have a latinesque flair.

Simpson depicted Thomashuxleya as being approximately proportioned like a big, robust dog (about 1.3 m in total length) with a relatively large head and obvious upper and lower canines. Vizcano et al. (2012) (citing unpublished PhD work by D. A. Croft) provide a mass estimate of 113 kg for Thomashuxleya, with c. 50 kg being given for the other Casamayoran isotemnids Anisotemnus and Pampatemnus.

However, Simpson’s aim to produce an isotemnid reconstruction seems to have been biased by his opinion that all taxa referred to this group were much alike: he considered it acceptable to combine elements from three different taxa in this reconstruction. In fact, it includes elements from Thomashuxleya externa, an unidentified species later referred to Anisotemnus distentus by Shockey & Flynn (2007), and a then-unnamed species that Simpson later named Pleurostylodon simulis.

Incomplete skull of Pleurostylodon modicus, from Simpson (1967).

It turns out that these three are actually quite different, most notably in forelimb proportions and joint morphology. In ulnar shape, Thomashuxleya is more like cursorial carnivorans and ungulates than are either Anisotemnus and Pleurostylodon. In Thomashuxleya, the ulna’s posterior border is concave and the olecranon process (the projecting flange that forms the bony lump of the ‘elbow’) is downcurved. In Anisotemnus and Pleurostylodon, the ulna’s posterior border is convex and the olecranon process curves upwards.

These differences suggest a more erect forelimb posture for Thomashuxleya, and a more flexed – or ‘crouched’ – posture in Anisotemnus and Pleurostylodon. These inferences are further supported by evidence from the distal end of the humerus (Shockey & Flynn 2007). Anisotemnus at least has a hand that appears suited for plantigrady, something you might expect for an animal with a flexed forelimb posture. Thomashuxleya was reconstructed as digitigrade by Simpson (1967) but we don’t know whether this really was the case. Maybe Thomashuxleya had adapted to browsing higher up than those other taxa, or maybe it was more specialised for a fully terrestrial life that didn’t involve any climbing or clambering – I’m speculating, but some aspects of ulnar shape seen in those other isotemnids recall features seen in climbing mammals, so maybe these animals did a bit of climbing and clambering on low branches and such.

Isotemnid ankle bones reveal that the amount of foot and ankle flexibility was quite low – the foot couldn’t be flexed or extended a great deal. An isolated foot (it can’t be referred to any specific isotemnid taxon but perhaps comes from a Pleurostylodon-like form) is pentadactyl with evidence for a divergent hallux. The metatarsals show that the animal – and presumably all isotemnids – had plantigrade hindfeet (Shockey & Flynn 2007). The general conclusion from ankle and foot anatomy here is that these animals couldn’t run and were ‘ambulatory’: walking animals that never really moved that fast, and never used a suspended phase when moving quickly. The caveat here is that both fast movement and running might still have been possible, however, since you can still do those things with inflexible ankles and non-cursorial limbs (Hutchinson et al. 2003).

Speculative life reconstruction of Thomashuxleya externa, by Naish.

In overall proportions, posture and abilities, Thomashuxleya was probably about similar to other robust-limbed mammals of the Paleogene – the pantodonts, astrapotheres and so on. All of these sorts of mammals probably used their robust, sometimes proportionally big, skulls, low-crowned cheek teeth and often big canines to forage among understorey plants for fruits, protein-rich leaves, fungi and so on. There aren’t precise analogues of these sorts of animals in the modern world (forest-dwelling rhinos and hippos come closest), but then, the thickly forested, mostly tropical Paleogene world, with its vast, closed-canopy forests and absence of pursuit predators was quite different from the Neogene one.

Typical isotemnid characters include spatulate incisors and low-crowned cheek teeth, but it has proved hard to find characters that reliably distinguish them from other toxodonts. Simpson (1967) and Cifelli (1993) pointed to the presence of an accessory crest on the lower molars as a possible isotemnid synapomorphy, but even this is known to be variable within the group and it seems to be lacking in some of them, including Thomashuxleya externa (Shockey et al. 2012). It's also present in other notoungulates (Shockey & Flynn 2007).

Besides Thomashuxleya, Anisotemnus and Pleurostylodon, who are the other isotemnids? Well, there’s Isotemnus, Hedralophus, Pampatemnus, Coelostynodon, Periphragnis, Rhyphodon, Distylophorus, Lophocoelus and Pleurocoelodon (McKenna & Bell 1997). Many of these taxa were named for maxillary or lower jaw fragments, with the key differences being small details of the cusps and associated structures, and the proportions of the teeth relative to one another.

Isotemnids within the toxodont radiation

Isotemnids have generally been considered the sister-group of remaining toxodonts – dubbed the ‘advanced Toxodontia’ by Cifelli (1993) – largely because they lack the derived characters that unite those other lineages (Simpson 1967, Cifelli 1993, Shockey et al. 2012). Furthermore, isotemnids possess a skeletal and dental anatomy that looks suitably ‘ancestral’ compared to what we see in ‘advanced Toxodontia’.

Hypothesised phylogeny of toxodont groups, showing different hand anatomies. Pentadactyl isotemnids are outside the clade in which digits I and V became reduced, and eventually lost. From Shockey et al. (2012).

Isotemnids have broad, pentadactyl hands and feet with blunt hooves, and a ‘primitive’ sort of scapula (where the acromial and metacromial processes overlie the shoulder joint), for example. However, the aberrant, clawed homalodotheriids (or homalodotheres, if you prefer) – we’ll get to them later – were often considered close to isotemnids in the past and indeed Ameghino united both in a group called Entelonychia. You might have guessed that this stems from Ameghino’s idea that these animals were close relatives of chalicotheres. Isotemnids and homalodotheriids were recovered as especially close relatives in some of Cifelli’s (1993) and Shockey et al.’s (2012) trees, and leontiniids (more on them later) group with both of these lineages when postcranial data is excluded from analysis (Shockey et al. 2012). The lesson here is probably that all three groups are conservatively similar in some aspects of morphology, not that they’re necessarily all closer to one another than they are to ‘advanced Toxodontia’.

Simpson (1967) described how isotemnid fossils are among the most commonly found of mammal remains at many sites. They were also among the largest of mammals present in the respective faunas, though it should be noted that Thomashuxleya is much larger than other Casamayoran isotemnid taxa.

And here ends the Tet Zoo guide to isotemnid toxodonts. Isotemnids are nice, but they’re arguably less interesting that the remaining toxodonts. We’ll be looking at those other forms soon enough.

Incidentally, this is Tet Zoo ver 3’s 99th article. What comes next?

For previous Tet Zoo articles on South American Cenozoic megafauna, see...

Refs - -

Cifelli, R. L. 1993. The phylogeny of the native South American ungulates. In Szalay, F. S., Novacek, M. J. & McKenna, M. C. (eds) Mammal Phylogeny: Placentals. Springer-Verlag (New York), pp. 195-216.

Croft, D. A. 1999. Placentals: endemic South American ungulates. In Encyclopedia of Paleontology, Fitzroy Dearborn Publishers (London), pp. 890-906.

Gabbert, S. L. 2004. The basicranial and posterior cranial anatomy of the families of the Toxodontia. Bulletin of the American Museum of Natural History 285, 177-190.

Hutchinson, J. R., Famini, D., Lair, R. & Kram, R. 2003. Are fast-moving elephants really running? Nature 422, 493-494.

McKenna, M. C. & Bell, S. K. 1997. Classification of Mammals: Above the Species Level. Columbia University Press (New York).

Patterson, B. D. & Costa, L. P. 2012. Bones, Clones & Biomes: The History and Geography of Recent Neotropical Mammals. The University of Chicago Press (Chicago and London).

Savage, R. J. G. & Long, M. R. 1986. Mammal Evolution: An Illustrated Guide. Facts on File Publications (New York & Oxford).

Shockey, B. J. & Flynn, J. J. 2007. Morphological diversity in the postcranial skeleton of Casamayoran (?Middle to Late Eocene) Notoungulata and foot posture in notoungulates. American Museum Novitates 3601, 1-26.

Shockey, B. J., Flynn, J. J., Croft, D. A., Gans, P.. Wyss, A. R. (2012). New leontiniid Notoungulata (Mammalia) from Chile and Argentina: comparative anatomy, character analysis, and phylogenetic hypotheses American Museum Novitates, 3737, 1-64 DOI: 10.1206/3737.2

Simpson, G. G. 1967. The beginning of the age of mammals in South America. Part II. Bulletin of the American Museum of Natural History 137, 1-260.

- . 1980. Splendid Isolation, the Curious History of South American Mammals. Yale University Press, New Haven, CT.

Vizcano, S. F., Cassini, G. H., Toledo, N. & Bargo, M. S. 2012. On the evolution of large size in mammalian herbivores of Cenozoic faunas of southern South America. In Patterson, B. D. & Costa, L. P. (eds) Bones, Clones & Biomes: The History and Geography of Recent Neotropical Mammals. The University of Chicago Press (Chicago & London), pp. 76-101.