Skip to main content

Marsupial `dogs', `bears', `sabre-tooths' and `weasels' of island South America: meet the borhyaenoids

This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American


I’ve decided to republish – in slightly updated form – the borhyaenoid text posted on Tet Zoo ver 2 back in July 2008. The text was originally published as three separate articles. It makes more sense to have it all together in the same place, so here are all those articles combined.

Even to novices with no special interest in the extinct wildlife of the Cenozoic, it should be obvious that ancient South America had what we might technically call a Really Awesome Faunal Assemblage. Astrapotheres, sebecosuchians, phorusrhacids, teratorns, gigantic caimans, madtsoiid snakes, sloths, glyptodonts… and this is only half of it, there’s so much more. It seems only fair and proper that yet another of South America’s extinct Cenozoic groups gets appropriate coverage on Tet Zoo, and yet again it’s a bunch of animals on which comparatively little information is freely available: the borhyaenoids, a long-lived and diverse group of carnivorous marsupials (or are they?). And it’s a particularly good time to bring this group to wider attention because, in a series of excellent recent papers, one worker has, almost single-handedly, done the most amazing job of bringing these animals to life.


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


Borhyaenoids were first recognised during the 1880s when Florentino Ameghino (1854-1911) described some new Patagonian fossil mammal remains dating to the Miocene. They were carnivores of some sort, with jaws and teeth reminiscent of those of carnivorous or omnivorous mammals like dogs or raccoons. He grouped them together in the newly recognised ‘family’ Acyonidae Ameghino, 1889 and suggested that they descended from opossums and included the ancestors of canids. That's right – Ameghino thought that the similarities present between South American mammals and their Northern Hemisphere counterparts were not the results of convergent evolution, but indicative of actual evolutionary relationships (Simpson 1980). Another name that Ameghino coined later on (in 1894) – Borhyaenidae – ended up superseding Acyonidae, and the name we’re using for the group here (Borhyaenoidea) is based on that one*. Borhyaenoids have also been called sparassodonts (for Sparassodonta Ameghino, 1894), and while this name is used by some recent authors it hasn’t been used as much. They’ve also been called Borhyaeniformes Szalay, 1962 and Borhyaenomorphia Archer, 1984.

* Marshall et al. (1978) petitioned the ICZN to suppress Acyonidae, and asked, while they were at it, that the spelling Borhyaena Ameghino, 1889 be formally recognised over Boryhaena [sic], the name that Ameghino originally used – apparently accidentally – in 1887.

Borhyaenoids within the marsupial radiation… or not

Most borhyaenoids have been characterised as ‘dog-like marsupials’, and if any common name is used for the group it’s this one. As we’ll see, at least some borhyaenoids were superficially dog-like, but others certainly weren’t, and if you want to find a more zoologically accurate way of describing these animals it might be to think of them as giant predatory opossums. It used to be thought that the dog-like marsupials of Australasia, the thylacines, were close kin of borhyaenoids, but anatomical and molecular data shows that this is not the case: thylacines are part of the exclusively Australasian marsupial clade that includes the dasyures. Borhyaenoids have also been considered allied to the stagodontids, a Cretaceous group of mostly North American omnivores or carnivores (some of which had bulbous teeth that look suited for crushing or breaking hard objects), to the deltatheroidans of Cretaceous Asia, and to the pediomyids of Cretaceous North America.

An extensive literature discusses these various proposals, most of it focusing on detailed characters of the braincase and ear region. Muizon et al. (1997) and Muizon (1999) showed that the stagodontid, deltatheroidan and pediomyid proposals were not supported by good character evidence, and argued that borhyaenoids should be united with opossums (didelphids) and their relatives in the clade Didelphimorphia. The key characters that unite these animals are found in the back of the skull.

If borhyaenoids are part of Didelphimorphia, then they’re part of the marsupial crown-group and can properly be considered part of Marsupialia. If they’re not part of Didelphimorphia, the competing hypothesis puts them outside the marsupial crown-group, meaning that they're stem-marsupials, or non-marsupialian metatherians. Note that Metatheria is more inclusive than Marsupialia, and that the name Marsupialia is restricted to the crown-group (e.g., Rougier et al. 1998, Asher et al. 2004). In a large and inclusive analysis of metatherians and other mammals, Rougier et al. (1998) found borhyaenoids and other alleged stem-didelphimorphians to be outside of Marsupialia; Asher et al. (2004) found a similar result for the supposed early borhyaenoid Mayulestes.

Let’s suppose, however, that borhyaenoids are didelphimorphians for a moment. While it might be difficult to imagine that a group of small, superficially rat-like predators are close kin of one of the most spectacular dynasties of big-bodied mammalian macropredators, note that Palaeocene proto-opossums, like Pucadelphys, are really very similar to taxa regarded by some as the most basal of borhyaenoids, like Mayulestes [Pucadelphys shown in adjacent image, from Paleocene Mammals of the World, but original is © MNHN]. The characters that distinguish borhyaenoids from other didelphimorphians are mostly minor cranial details like the absence of prootic canal, the form of the tympanic sinus, and the presence of nasal bones that don’t overhang the narial opening. Borhyaenoids are also characterised by specialised shearing teeth (Muizon 1999).

Were borhyaenoids pouched? That’s not a dumb question, as pouches have been repeatedly lost within metatherian evolution, and particularly within opossums. Articulated borhyaenoid skeletons lack the epipubic bones used in other metatherians to support the pouch, so pouches might have been absent. However, some marsupials (like Thylacinus) that lack ossified epipubic bones have cartilaginous ones, so we can’t be sure.

While people have often made proposals about how certain borhyaenoids lived and hunted, detailed comparative work that analyses the anatomical details of these animals and compares them with those of living carnivorous mammals has generally been lacking. Indeed, harking back to comments I made about the CEE Functional Anatomy meeting, the widely held view that this sort of work is old-fashioned, outdated and redundant ignores the massively under-appreciated fact that, in an absurdly high number of cases, it’s never been done at all. Thanks almost entirely to one worker – Christine Argot of the Muséum National d’Histoire Naturelle (Paris) – we now have a very robust, incredibly detailed literature on borhyaenoid functional anatomy. Argot’s papers contain so much information that I’m only going to be reviewing but a fraction of the areas she covers. Given that lengthy papers on functional anatomy of any type are pretty rare, it cannot be over-emphasised how important studies like this are. What’s also of special interest as goes borhyaenoids is that Argot’s work has shown how various taxa, repeatedly depicted in the popular and semi-technical literature, seem to have been quite different from their ‘conventional’ image.

Mayulestes and Allqokirus

Mayulestes ferox from the Santa Lucía Formation of Tiupampa, Bolivia, has typically been regarded as the sister-taxon to the rest of the borhyaenoid radiation (Muizon 1994). However, Rougier et al. (1998) didn’t find Mayulestes to be part of Borhyaenoidea, and this was followed by Forasiepi et al. (2006). Tiupampa gives its name to the Tiupampan Land Mammal Age, a time span dated to the early Palaeocene.

A second Tiupampan borhyaenoid, Allqokirus australis, has been regarded as close to Mayulestes and both have been regarded as part of the ‘family’ Mayulestidae Muizon, 1994. The fact that Allqokirus is known only from teeth makes this difficult to verify, however. Mayulestes was weasel-sized and weighed less than 1 kg: probably, in fact, less than 500 g. By comparing its pelvis and hindlimb with those of extant marsupials, Argot (2002) showed that this animal had a mobile lumbar region and was agile and well able to climb and move with agility across uneven surfaces. However, the forelimb has features more suggestive of a climbing lifestyle (Argot 2001). It wasn’t a cursorial animal, running around in open habitats, nor was it good at leaping. It should probably be imagined as scansorial: as a predominantly terrestrial denizen of cluttered, three-dimensional habitats like tangled forest floors, but well able to climb.

The Tiupampan fauna includes an assortment of opossums and mioclaenids, and there are also some real oddballs in the fauna, like the pantodont Alcidedorbignya and the opossum-like Szalinia and Jaskhadelphys. It’s been hypothesised that Mayulestes preyed on frogs, small marsupials, and the eggs of crocodilians, turtles and birds (Argot 2004a). As the potential ‘most basal’ member of Borhyaenoidea, Mayulestes might show us what the ancestral lifestyle and morphology of the group was. However, as we’ve just seen, there is now doubt as to whether Mayulestes is definitely a borhyaenoid.

With Mayulestes and Allqokirus out of the way, the remaining borhyaenoids have been recognised as a clade (unnamed, so far as I know) united on the basis of braincase details, palatal morphology and an incisor compliment reduced to four uppers and three lowers* (Muizon 1994, 1999). Conventionally, these ‘higher borhyaenoids’ have been grouped into five ‘families’: Hathlyacynidae, Prothylacynidae, Proborhyaenidae, Borhyaenidae and Thylacosmilidae. These have also been regarded as ‘subfamilies’ in those classification schemes where the group here termed Borhyaenoidea has been regarded as a ‘family’. Of these groups, hathlyacynids (which differ from other borhyaenoids in possessing a pneumatised squamosal) appear to form the sister-taxon to a clade that includes the remaining taxa (Muizon 1994, 1999).

* The primitive condition for marsupials is five upper incisors and four lowers.

Long-snouted marsupial martens and false thylacines

Hathlyacynidae is the biggest and longest-lived borhyaenoid clade, with members that range in age from Late Palaeocene (Patene) to Early Pliocene (Notocynus and Borhyaenidium). Notocynus might even have made it to the Late Pliocene, but this is debatable. Incidentally, there are two ways of spelling the name of this group, with some authors spelling it Hathliacynidae. Most of the 17 or so hathlyacynid genera are known only from jaw fragments, and these show that members of the group were conservative over their long history: in fact Marshall (1981) described hathlyacynid genera as “monotonously alike”.

While often characterised as dog-like, those taxa known from good remains show that at least some hathlyacynids were marten-like predators with good climbing abilities. So far as we know they were all long snouted, with short gaps in the toothrow both between the premolars, and between the first premolar and the canine. Their teeth suggest that they were carnivorous or even omnivorous generalists, though there is a trend for later members of the group to have more specialised carnassials (Marshall 1981). Two hathlyacynids in particular, Sipalocyon gracilis and Cladosictis patagonica (both from the Santacrucian beds of Early Miocene Patagonia), have been subjected to detailed functional analyses.

Based on its divergent hallux and semi-opposable pollex, Sinclair (1906) thought that Sipalocyon was plantigrade and probably arboreal, and comparison of its bones with those of extant mammals suggest that it was probably scansorial (Argot 2003a, 2004a). Its mobile thumb suggests that it was good at grabbing and manipulating prey. Cladosictis has been repeatedly depicted in the popular and semi-technical literature as an amphibious, otter-like predator. This idea appears to come from Savage’s (1977) statement that the forelimbs of Cladosictis recall those of an otter in their proportions. However, otter-like forelimb proportions do not mean an otter-like lifestyle, and in fact the forelimb proportions of otters aren’t really reliably different from those of other (non-swimming) mustelids.

Cladosictis was about 20-25 cm tall at the shoulder and weighed 4-8 kg; it was relatively short-legged with limb details suggesting a good climbing ability [adjacent skeletal reconstruction from Argot 2004a. Image © C. Argot. Scale bar = 10 cm]. It also had a semi-opposable thumb, and hence was able to grasp branches and manipulate prey. The closest analogue is perhaps the Tayra Eira barbata, a tropical American mustelid that forages terrestrially but is also a good climber (Argot 2003a, 2004a). Based on contemporary fossils and morphology, Cladosictis probably predated on small mammals, birds, reptiles and frogs. The neck was about similar in length to that of a modern canid (c. 37% of thoracolumbar length); large ventral processes on the neck vertebrae demonstrate that the musculature here was powerful. Reasonably long and powerfully muscled necks of this sort are typical of borhyaenoids.

All the other hathlyacynids

Do we know if other hathlyacynids were doing the same sort of thing as Sipalocyon and Cladosictis? Unfortunately we don’t know enough about them to be sure. The forelimb bones of little Pseudonotictis pusillus (also from the Santacrucian) indicate climbing abilities, but this taxon had particularly slender bones, suggesting that it was doing something unusual… but just what we’re not sure (Argot 2003a). The lower jaw of P. pusillus is about 4 cm long, suggesting a weasel-like size of less than 40 cm. It’s the smallest hathlyacynid known, but its probable close relative Notictis ortizi from the Huayquerian (= Late Miocene) of Argentina was nearly as small (Marshall 1981). Marshall (1981) and Villarroel & Marshall (1983) further proposed that Santacrucian Perathereutes pungens shared an ancestor with Pseudonotictis, and that Perathereutes was close to the ancestry of Borhyaenidium from the Huayquerian. In both Perathereutes and Sipalocyon the first lower premolar is set at an angle relative to the other teeth, leading Marshall (1981) to suggest that both shared an ancestor. Marshall (1981) also regarded Notocynus hermosicus from the Montehermosan (= late Late Miocene) as a close relative or descendant of Sipalocyon as the two share tooth cusp characters.

However, Sipalocyon was regarded as closely related to Notogale by Forasiepi et al. (2006), as both possess a transverse canal in the auditory region. This character is also present in Cladosictis, however (Muizon 1999). Santacrucian and Laventan (= Middle Miocene) Acyon – resurrected from the synonymy of Anatherium by Forasiepi et al. (2006)* – shares a single tooth character with Cladosictis and might be closely related to it, and Chasiocostylus castroi is also similar and probably close to Cladosictis.

* The type species of Anatherium, A. defossus, was referred to Cladosictis sp. by Forasiepi et al. (2006).

Sallacyon hoffstetteri has been regarded as a basal hathlyacynid (Muizon 1999, Forasiepi et al. 2006), as has Patene (Marshall 1981). Procladosictis anomala from the Mustersan (= middle Eocene) resembles Patene. Several other genera, including Agustylus, Ictioborus, and Amphithereutes, continue to be listed by some authors (e.g., McKenna & Bell, 1997), but are regarded as synonymous with other taxa by others (e.g,. Marshall 1981). Palaeocladosictis mosesi, sometimes listed as a hathlyacinid, was based on an ungulate tooth according to Marshall (1978).

Prothylacinids: climbing thylacine-hyaena-binturong hybrids

Character evidence indicates that hathlyacynids are the sister-group to a borhyaenoid clade that includes all the remaining taxa. Among these ‘remaining taxa’, we begin with the prothylacinids (or prothylacynines): an Oligocene-early Pliocene group regarded by some as part of Borhyaenidae: both groups share characters not seen in other borhyaenoids (Muizon 1999). However, prothylacinids and borhyaenids are more usually imagined as sister-groups, which, if true, means that prothylacynids must have a ghost-lineage going back to the Early Eocene at least (as there are apparently borhyaenids this old, like Argyrolestes and Angelocabrerus).

Lycopsis from the Santacrucian and Friasian (= late Early Miocene) has conventionally been regarded as one of the most basal prothylacinids (Marshall 1979). However, more recent studies have failed to recover this position, and have instead found it to be the sister-taxon of, or basal member of, a ((prothylacinid + borhyaenid) + (proborhyaenid + thylacosmilid)) clade (Muizon 1999, Babot et al. 2002, Forasiepi et al. 2006). Again, the characters involved are all trivial details of the back of the skull and if I explain them I can see myself adding 1000 words to this article.

Thanks to a wonderfully near-complete skeleton of L. longirostrus* from the Friasian of Colombia [shown above] we have a reasonably good idea of the morphology and functional anatomy of this animal (Marshall 1977a, Argot 2004b). About 35 cm tall at the shoulder and weighing about 15 kg (Argot 2004a), it possesses an unusual combination of features. The relatively elongate, straight-boned forelimb and semi-digitigrade manus indicate a trend towards cursoriality, but the semi-opposable thumb, short metatarsals and prominent hallux are at odds with this. Lycopsis doesn’t appear to have been a particularly good climber, but it could probably clamber in trees if need be, and Argot (2004b) noted that this might have been necessary given the densely forested environment it inhabited and the presence of a diverse crocodilian assemblage. Perhaps it was an ambush predator of small and mid-sized prey, able to engage in some running and some climbing, but not specialised for either. While previously imagined as thylacine-like in lifestyle, it probably mostly preyed on small rodents, and indeed we know it did at least sometimes, as rodent bones and a tooth were preserved adjacent to the pelvis of one specimen (Marshall 1977a).

* Argot (2004b) mistakenly spelt it L. longirostris [sic].

Moving on to prothylacinids proper, Argot (2003b) showed that Prothylacinus patagonicus from the Santacrucian had a powerful neck, short, muscular limbs and a flexible body. Its manual phalanges were proportionally long, its manual unguals were deep, sharply curved and with large flexor tubercles, and it probably had a fairly large, pseudo-opposable pollex.

What we know of its tail indicates that the organ was long (incorporating 20-30 vertebrae) and gradually tapering. Its metapodials were proportionally short and its feet were plantigrade, with a long first metatarsal probably supporting a long plantar pad that would have enabled the foot to grip curved surfaces.

Combined, these features strongly suggest that Prothylacinus was another scansorial borhyaenoid: a powerful, agile climber that used flexible hands in gripping branches and a muscular neck and large skull to support the weight of large prey items. No living carnivorous mammal is exactly like this, but the strongest similarity is with binturongs Arctictis (Argot 2003b) [adjacent binturong by Tassilo Rau]. Binturongs are large, slow, agile climbers, with robust, muscular limbs and long tails. They’re predominantly frugivorous, so (so far as we know) are unlike Prothylacinus in this respect. Argot (2004a) speculated that the prey of Prothylacinus might have included rodents, caenolestoids and some of the smaller sloths. Incidentally, Prothylacinus was relatively large, with an estimated weight of 27-37 kg. However, such a size is not at all incompatible with a scansorial lifestyle, as demonstrated by living wolverines, sun bears, clouded leopards and leopards. Prothylacinus was from the Santacrucian fauna, so did it complete with the also scansorial Cladosictis and Sipalocyon? The difference in size between the hathlyacynids and prothylacinid indicate that they occupied different niches.

The crowded cheek teeth, shortened premolars and molars and tightly fused mandibular symphysis indicate that Prothylacinus was a specialised carnivore less suited for omnivory than other prothylacinids. Indeed, scars indicating large masseter muscles, combined with evidence from tooth cusp morphology, have led other prothylacinids (like Montehermosan Stylocynus paranensis: shown here, from Marshall 1979) to be regarded as omnivores with a bear-like diet (Marshall 1979).

The largest prothylacynid is Dukecynus magnus from La Venta in Colombia. Charactised by a particularly narrow, elongate rostrum (Goin 1997), its teeth suggest that it was more carnivorous than some of the other taxa, and its molars suggest close affinities with Pseudolycopsis cabrerai from the Chasicoan. Pseudothylacinus from the Colhuehuapian was thought by Marshall (1979) to be close to the ancestry of Pseudolycopsis.

The not so bear-like borhyaenids

Borhyaenids (yes, borhyaenid borhyaenoids) include about ten genera of superficially dog- or thylacine-like borhyaenoids. The oldest (Nemolestes) is from the Early Eocene or possibly Late Palaeocene while the youngest (Eutemnodus) is from the Early Pliocene (Marshall 1978)*. Easily the best known member of the group is Borhyaena tuberata from the Santacrucian (= Early Miocene). Like the contemporary Cladosictis, it has frequently been depicted in the popular and semi-technical literature. As is also the case with Cladosictis, its ‘conventional’ image is – so it turns out – pretty far off the mark.

* An alleged borhyaenid from the Late Miocene or Early Pliocene – Parahyaenodon argentinus from Argentina – was reidentified as a procyonid (a member of the raccoon family) by Forasiepi et al. (2007).

The best known life restoration of this animal (that produced by John Long for Mammal Evolution: An Illustrated Guide) depicts it as something like a long-tailed bear, and I can recall thinking that Borhyaena was giant; perhaps similar in size to a brown bear. In fact, weight estimates (Argot 2003b) put it at 19-29 kg, which is about equivalent to a small hyaena or wolf, or a large thylacine. However, its robust skull, with its particularly broad zygomatic arches, indicate a disproportionately large amount of skull and neck musculature, and mass estimates for the whole animal based on skull proportions alone (and assuming body proportions resembling those of extant carnivorous mammals) give Borhyaena an inflated mass of 74 kg, a vivid illustration of why knowing the overall proportions of an animal are important when estimating its weight (Van Valkenburgh 1985, 1987). The substantial neck musculature suggests an ability to carry heavy loads.

A semi- or fully digitigrade manus, short, blunt claws, and forelimbs that were restricted to parasagittal movement and exhibit reduced distal musculature indicate that Borhyaena was terrestrial and cursorial – in fact, the most cursorial of all borhyaenoids. However, its limbs weren’t as proportionally elongate as those of extant cursorial predators. That might not have been such a problem, because virtually all the potential large-bodied prey (a assortment of xenarthrans, astrapotheres, notoungulates and big rodents) were not cursorial either: in the Santacrucian fauna, only proterotheriid and macraucheniid litopterns can be considered cursorial (Argot 2004a). While previously depicted as being plantigrade, what is known of ankle morphology and forelimb proportions show that Borhyaena was more likely digitigrade in the hindlimb (Argot 2003b) [life restoration and skeletal reconstruction above © C. Argot; skull photo below by Ghedoghedo].

Whether the other borhyaenids were like ‘new-look Borhyaena’ remains to be shown. Fredszalaya hunteri from the Deseadan (Late Oligocene) of Bolivia was suggested by Shockey & Anaya (2008) to be closely related to Borhyaena on the basis of the form of the alisphenoid, but the unreduced shelves and cusps on its molars suggest that it was omnivorous while the shape of its calcaneum suggest possible climbing abilities (Shockey & Anaya 2008). Fredszalaya is smaller than Borhyaena and with a much shorter snout.

Proborhyaenids, the ‘marsupial bears’

Proborhyaenids have usually been thought of as as a group of hyaena-like or bear-like borhyaenoids. Marshall (1977b) thought that proborhyaenids might be the most basal of borhyaenoids (while at the same time the most specialised), but it has more recently been argued that proborhyaenids share derived characters with the sabre-toothed cat-like thylacosmilids (Muizon 1994, 1999, Babot et al. 2002). One of the most interesting characters present in both proborhyaenids and thylacosmilids is the presence of ever-growing, open-rooted upper canines. Both groups also exhibit strongly projecting occipital condyles, indicating that they were capable of greater rotation and movement at the head-neck joint than other borhyaenoids. Babot et al. (2002) suggested that the presence of fine ridges and grooves on the canine roots might be diagnostic for proborhyaenids, but the fact that thylacosmilids have such strongly modified canines raises the possibility that thylacosmilids had this character ancestrally but later reversed it. Babot (2005) found Proborhyaenidae of tradition to be paraphyletic, but I haven’t seen this (unpublished) study and don’t know the details.

Proborhyaenids are unique to the Eocene and Oligocene: there are only four recognised genera (Callistoe, Arminiheringia*, Paraborhyaena and Proborhyaena). Proborhyaena in particular has often been referred to as huge and bear-like, with there being estimates here and there of skulls perhaps 60 cm long [P. gigantea holotype lower jaw shown here, from Patterson & Marshall (1978). From top to bottom, the specimen is shown in labial, occlusal and lingual views. Scale bar = 50 mm].

However, as has turned out to be the case for other archaic carnivorous mammals (example: hyaenodontids), big, bear-sized skulls do not necessarily mean big, bear-sized bodies, and an articulated proborhyaenid skeleton (read on) shows that these animals had big heads for their size.

* Arminiheringia was originally given its own ‘family’, Arminiheringiidae Amegino, 1902.

As is the case with most borhyaenoid genera, proborhyaenids are mostly known from fragmentary jaws and other parts of skulls. The 2002 discovery of a fantastic, near-complete proborhyaenid skeleton was thus a major boon to our understanding of these animals. The skeleton [shown above] represents the new taxon Callistoe vincei and is from the Casamayoran (= Early Eocene) Lumbrera Formation of Salta, Argentina (Babot et al. 2002). The generic name comes from Callisto, the Arcadian nymph loved by Zeus and changed by him into a bear to protect her from Hera’s wrath (Jupiter has a moon called Callisto).

In contrast to other proborhyaenids, Callistoe has a gracile, narrow snout and might have superficially resembled a thylacine in facial shape. In overall proportions, it most resembles marten-like mustelids and red pandas (Argot & Babot 2011): unlike those animals, however, it lacks limb and vertebral characters associated with climbing. Instead, the form of its limb joints, the presence of an ossified patella and the shapes of its limb bones show that it was restricted to a parasagittal gait, possessed little flexibility in its elbow, wrist, knee and ankle, and hence was specialised for terrestrial walking and running. Argot & Babot (2011) described how its very long, gently curved manual claws – totally unlike those of other borhyaenoids – suggest a digging ability. Intriguingly, what appears to be a healed fracture on one digit suggests 'heavy use' of its hands. Meanwhile, the slender outer toes on its very short hindfeet indicate cursoriality. It probably weighed about 23 kg. Callistoe lived alongside a variety of armadillos, rodents and small notoungulates, so perhaps it needed to cover large distances on foot before digging for such prey when it discovered them. Remember that this animal lived during the Early Eocene: it long pre-dated the digging mustelids and other such placental carnivores than evolved elsewhere.

Arminiheringia had unusual procumbent lower canines: there has actually been some disagreement as to whether this is natural or not (Bond & Pascual 1983), but it does appear to be according to Babot et al. (2002) (A. auceta shown here, from Simpson 1932). Paraborhyaena also had somewhat procumbent lower canines and, unlike other members of the group, it had only a single pair of lower incisors (Shockey & Anaya 2008) – a character seen elsewhere in thylacosmilids. After Thylacosmilus, Simpson (1932) regarded Arminiheringia auceta (one of three species in the genus) as the most specialised borhyaenoid, and also one of the largest, “about the size of the great Pharsophorus lacerans” (Pharsophorus is a borhyaenid). Exactly how big Pharsophorus was, however, I have no idea – and was it supposed to be bigger than, say, Proborhyaena?

The marsupial sabre-tooths

Finally, we come to the superficially cat-like thylacosmilids, the only borhyaenoid group that are at all familiar, and all thanks to the incredible Thylacosmilus atrox Riggs, 1933 of the Late Miocene and Early Pliocene of Catamarca, north-western Argentina. In contrast to other borhyaenoids, thylacosmilids were short-faced and some of them, at least, had a complete bony postorbital bar. T. atrox is the only thylacosmilid we ever hear about, but it isn’t the only one.

Achlysictis Ameghino, 1891, known only from mandibular fragments and teeth (three species have been named), was smaller than T. atrox and differed from it in tooth cusp characters. Similar comments can be made about Hyaenodonops Ameghino, 1908, known only from teeth (though postcrania has been referred to it). Notosmilus Kraglievich, 1960, named for a maxilla and canine, was only about half the size of T. atrox. Marshall (1976) regarded all of these taxa as distinct, but Goin & Pascual (1987) and McKenna & Bell (1997) regarded them all as synonyms of Thylacosmilus. Of course, if that’s true, then the generic name Thylacosmilus is pre-dated by both Achlysictis and Notosmilus and the ICZN would have to be petitioned if we want to preserve Thylacosmilus (which we do). Incidentally, Riggs (1933) named a second species of Thylacosmilus, T. lentis, but this was argued by Marshall (1976) to be synonymous with T. atrox.

Most recently, Goin (1997) named the Middle Miocene species Anachlysictis gracilis from the La Venta site in Colombia. This animal differed from the other thylacosmilids in smaller size, in possessing a flattened skull roof, and in lacking a postorbital bar. And another thylacosmilid, Patagosmilus goini, has since been named from the middle Miocene of Argentina (Forasiepi & Carlini 2010). The form of the one preserved hand bone and shape of the manual claw suggest that Patagosmilus was able to climb, but this is uncertain given the poor nature of the material. It otherwise seems to have been similar to Thylacosmilus.

You might think you know Thylacosmilus, but (as is so often the case) it’s rather stranger than usually thought. [Adjacent photo by Claire Houck]. Based on Riggs’s skull reconstruction, some authors (and artists) have concluded that it had no upper incisors at all, in which case you really have to wonder how it pulled tissue from prey after killing it. Such a configuration would be so unusual compared to other predatory mammals that Churcher (1985) cautioned against it, and it all rests on the fact that the critical regions of the premaxillary bones are absent. Furthermore, there’s enough space between the huge upper canines for small incisors, and there are wear facets on the single pair of lower incisors that can only have been caused by interaction with upper incisors (Churcher 1985).

The laterally compressed, ever-growing upper canines were not just rooted in maxillary sockets as they are in sabre-toothed cats, but arced up and over the orbits, forming a notably convex skull roof. The premolars and molars were narrow lineal blades, specialised for slicing. Huge, laterally flattened flanges – sometimes called genial flanges – grew downwards from the lower jaw, and the sabre teeth would have rested against their sides when the jaws were closed. Riggs (1934) illustrated these diverging laterally toward their tips, and several artists followed his lead. However, other authors have considered this inaccurate (Turnbull 1976, Marshall 1976, Churcher 1985), and the flanges were more likely fairly parallel.

Protuberances, rugosities and excrescenses at the back and base of the skull show that a substantial amount of musculature allowed both great power and fine control to be exerted over the head, and large muscle attachment sites on the neck vertebrae also show that the neck was very strong and flexible. Several studies have looked at these features: Marshall (1976) discussed skull function and its possible role in behaviour, Turnbull (1976) reconstructed the cranial musculature, and Turner & Antón (1997) and Argot (2004c) analysed postcranial morphology.

How exactly did thylacosmilids live? Again, Argot’s (2004c) detailed functional analysis provides us with a huge amount of information. With its semi-plantigrade hindfeet, stout fibulae and curved tibiae, Thylacosmilus looks poorly suited for fast running. A massively powerful upper arm and a reinforced and relatively inflexible lumbar region imply that it was an ambush predator that attacked prey after a short dash.

It’s now widely thought that sabre-toothed predatory mammals used their elongate and delicate weapons for precise attacks inflicted on the ventral surface of the neck, and for this to work the predator has to be able to physically manipulate and restrain the prey, and use coordinated and precise neck and head movements to attack in the right place. A semi-opposable thumb and tremendous forelimb strength suggest that Thylacosmilus could pin down and restrain prey, and its elongate, powerful neck demonstrates the presence of a ‘neck-driven’ precision biting style. In their study of bite forces seen in various mammalian predators, Wroe et al. (2005) found bite forces of Thylacosmilus to be very low. This is also the case for sabre-toothed cats like Smilodon, and it indicates that these animals were not ‘power biting’ like short-toothed predators, but driving their unusual killing style with their neck musculature. Most thylacosmilid reconstructions make the animals look far too cat-like, and also give them completely incorrect limb details [brilliant illustrations below by Carl Buell: be sure to check out - and 'like' - Carl Buell Illustration at facebook].

Here’s an interesting thing to consider within the context of thylacosmilid biology and behaviour. Given that sabre-toothed biting appears to have been specialised and difficult in actual, practical terms, it’s inferred that sabre-toothed predators have to go through a very dangerous ‘apprenticeship’ in which they learn how to successfully inflict a bite without getting smacked in the head or breaking a tooth. But, for this to occur, juveniles and adults must stay together for an extended post-weaning period. In marsupials* in general, this is rare, with juveniles rarely staying with their parents for more than a few weeks once weaning is finished. Were thylacosmilids a major exception? That is, did they have some sort of extended system of parental care? Unfortunately, we just don’t know.

* Remember that borhyaenoids may not actually be marsupials, but members of the more inclusive clade Metatheria.

What were thylacosmilids preying on anyhow? Argot (2004c) proposed that mesotheriid notoungulates, litopterns and big rodents like capybaras were among the prey of Thylacosmilus. And it’s also interesting to note that thylacosmilids were living alongside other borhyaenoids: in the Huayquerian, Thylacosmilus was sharing its habitat with late-surviving hathlyacinids, the probably omnivorous prothylacinid Stylocynus, and the borhyaenid Eutemnodus. Phorusrhacids were around too: did they compete with borhyaenoids, or did the groups occupy different niches and avoid competition? So many animals, so many questions.

One final point. I hope it's clear by now that there's quite a bit of very good technical literature on borhyaenoids. But what if you want semi-technical or popular stuff, preferably well-illustrated with nice photos and life restorations? Sad to say, we're still at the stage where such a work simply does not exist. This is despite the recent publication of some very good volumes on the fossil fauna of Cenozoic South America (e.g., Sánchez-Villagra et al.’s 2010 Urumaco & Venezuelan Paleontology and Patterson & Costa’s 2012 Bones, Clones & Biomes: The History and Geography of Recent Neotropical Mammals). Takehome point: there is still a gap in the market. We need a good, non-technical, well-illustrated volume on the Cenozoic fossil fauna of South America.

Many thanks to Christine Argot for her kind assistance with this article.

Refs – -

Argot, C. 2001. Functional-adaptive anatomy of the forelimb in the Didelphidae, and the paleobiology of the Paleocene marsupials Mayulestes ferox and Pucadelphys andinus. Journal of Morphology 247, 51-79.

- . 2002. Functional-adaptive analysis of the hindlimb anatomy of extant marsupials and the paleobiology of the Paleocene marsupials Mayulestes ferox and Pucadelphys andinus. Journal of Morphology 253, 76-108.

- . 2003a. Postcranial functional adaptations in the South American Miocene borhyaenoids (Mammalia,Metatheria): Cladosictis, Pseudonotictis and Sipalocyon. Alcheringa 27, 303-356.

- . 2003b. Functional adaptations of the postcranial skeleton of two Miocene borhyaenoids (Mammalia, Metatheria), Borhyaena and Prothylacinus, from South America. Palaeontology 46, 1213-1267.

- . 2004a. Evolution of South American mammalian predators (Borhyaenoidea): anatomical and palaeobiological implications. Zoological Journal of the Linnean Society 140, 487-521.

- . 2004b. Functional-adaptive analysis of the postcranial skeleton of a Laventan borhyaenoid, Lycopsis longirostris (Marsupialia, Mammalia). Journal of Vertebrate Paleontology 24, 689-708.

- . 2004c. Functional-adaptive features and palaeobiologic implications of the postcranial skeleton of the late Miocene sabretooth borhyaenoid Thylacosmilus atrox (Metatheria). Alcheringa 28, 229-266.

- . & Babot, J. 2011. Postcranial morphology, functional adaptations and palaeobiology of Callistoe vincei, a predaceous metatherian from the Eocene of Salta, north-western Argentina. Palaeontology 54, 411-480.

Asher RJ, Horovitz I, & Sánchez-Villagra MR (2004). First combined cladistic analysis of marsupial mammal interrelationships. Molecular phylogenetics and evolution, 33 (1), 240-50 PMID: 15324852

Babot, M. J. 2005. Los Borhyaenoidea (Mammalia, Metatheria) del Terciario inferior del Noroeste argentino. Aspectos filogenéticos, paleobiológicos y bioestratigráficos. Unpublished PhD thesis, Universidad Nacional de Tucumán, Tucumán, Argentina, 454 pp

- ., Powell, J. E. & de Muizon, C. 2002. Callistoe vincei, a new Proborhyaenidae (Borhyaenoidea, Metatheria, Mammalia) from the Early Eocene of Argentina. Geobios 35, 615-629.

Bond, M. & Pascual, R. 1983. Nuevos y elocuentes restos craneanos de Proborhyaena gigantea Ameghino, 1897 (Marsupialia, Borhyaenidae, Proborhyaeninae) de la Edad Deseadense. Un ejemplo de coevolución. Ameghiniana 20, 47-60.

Churcher, C. S. 1985. Dental functional morphology in the marsupial sabre-tooth Thylacosmilus atrox (Thylacosmilidae) compared to that of felid sabre-tooths. Australian Mammalogy 8, 201-220.

Forasiepi, A. M. & Carlini, A. A. 2010. A new thylacosmilid (Mammalia, Metatheria, Sparassodonta) from the Miocene of Patagonia, Argentina. Zootaxa 2552, 55-68.

- ., Martinelli, A. G. & Goin, F. J. 2007. Revisión taxonómica de Parahyaenodon argentinus Ameghino y sus implicancias en el conocimiento de los grandes mamíferos carnívoros del Mio-Plioceno de América de Sur. Ameghiniana 44, 143-159.

- ., Sánchez-Villagra, M. R., Goin, F. J., Takai, M., Shigehara, N. & Kay, R. F. 2006. A new species of Hathliacynidae (Metatheria, Sparassodonta) from the middle Miocene of Quebrada Honda, Bolivia. Journal of Vertebrate Paleontology 26, 670-684.

Goin, F. J. 1997. New clues for understanding Neogene marsupial radiations. In Kay, R. F., Madden, R. H., Cifelli, R. L. & Flynn, J. J. (eds) Vertebrate Paleontology in the Neotropics: The Miocene fauna of La Venta, Colombia. Smithsonian Institution Press (Washington, D.C.), pp. 187-206.

- . & Pascual, R. 1987. New on the biology and taxonomy of the marsupials Thylacosmilidae (late Tertiary of Argentina). Anales de la Academia Nacional de Ciencias Exactas, Físicas y Naturales 39, 219-246.

Marshall, L. G. 1976. Evolution of the Thylacosmilidae, extinct saber-tooth marsupials of South America. PaleoBios 23, 1-30.

- . 1977a. A new species of Lycopsis (Borhyaenidae: Marsupialia) from the La Venta fauna (Late Miocene) of Colombia, South America. Journal of Paleontology 51, 633-642.

- . 1977b. Cladistic analysis of borhyaneoid, dasyuroid, didelphoid, and thylacinid (Marsupialia: Mammalia) affinity. Systematic Zoology 26, 410-425.

- . 1978. Evolution of the Borhyaenidae, extinct South American predaceous marsupials. University of California Publications in Geological Sciences 117, 1-89.

- . 1979. Review of the Prothylacininae, an extinct subfamily of South American “dog-like” marsupials. Fieldiana Geology New Series 3, 1-50.

- . 1981. Review of the Hathlyacyninae, an extinct subfamily of South American “dog-like” marsupials. Fieldiana Geology New Series 7, 1-120.

- ., Clemens, A., Hoffstetter, R. J., Pascual, R., Patterson, B., Tedford, R. H. & Turnbull, W. D. 1978. Acyonidae Ameghino, 1889 (Mammalia): supplement to proposal to suppress this name. Bulletin of Zoological Nomenclature 35, 12-14.

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

Muizon, C. 1994. A new carnivorous marsupial from the Palaeocene of Bolivia and the problem of marsupial monophyly. Nature 370, 208-211.

- . 1999. Marsupial skulls from the Deseadan (late Oligocene) of Bolivia and phylogenetic analysis of the Borhyaenoidea (Marsupialia, Mammalia). Geobios 32, 483-509.

- ., Cifelli, R. L. & Paz, R. C. 1997. The origin of the dog-like borhyaenoid marsupials of South America. Nature 389, 486-489.

Patterson, B. & Marshall, L. G. 1978. The Deseadan, Early Oligocene, Marsupialia of South America. Fieldiana Geology 41, 37-100.

Riggs, E. S. 1933. Preliminary description of a new marsupial sabertooth from the Pliocene of Argentina. Geological Series of FieldMuseum of Natural History 6, 61-66.

- . 1934. A new marsupial saber-tooth from the Pliocene of Argentina and its relationships to other South American predacious marsupials. Transactions of the American Philosophical Society 24, 1–32.

Rougier, G. W., Wible, J. R. & Novacek, M. J. 1998. Implications of Deltatheridium specimens for early marsupial history. Nature 396, 459-463.

Savage, R. J. G. 1977. Evolution in carnivorous mammals. Palaeontology 20, 237-271.

Shockey, B. J. & Anaya, F. 2008. Postcranial osteology of mammals from Salla, Bolivia (Late Oligocene): form, function, and phylogenetic implications. In Sargis, E J. & Dagosto, M. (eds) Mammalian Evolutionary Morphology: A Tribute to Frederick S. Szalay. Springer Science (Dordrecht, The Netherlands), pp. 135-157.

Simpson, G. G. 1932. Skulls and brains of some mammals from the Notostylops beds of Patagonia. AmericanMuseum Novitates 578, 1-11.

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

Sinclair, W. J. 1906. Marsupialia of the Santa Cruz beds. In Scott, W. B. (ed) Reports of the PrincetonUniversity Expedition to Patagonia, 1896-1899. Princeton University, pp. 333-460.

Turnbull, W. D. 1976. Restoration of masticatory musculature of Thylacosmilus. In Churcher, C. S. (ed) Athlon, Essays in Palaeontology in Honour of Loris Shano Russell. Royal Ontario Museum (Toronto), pp. 169-185.

Turner, A. & Antón, M. 1997. The Big Cats and Their Fossil Relatives. Columbia University Press, New York.

Van Valkenburgh, B. 1985. Locomotor diversity within past and present guilds of large predatory mammals. Paleobiology 11, 406-428.

- . 1987. Skeletal indicators of locomotor behaviour in living and extinct carnivores. Journal of Vertebrate Paleontology 7, 162-182.

Villarroel, C. & Marshall, L. G. 1983. Two new late Tertiary marsupials (Hathlyacyninae and Sparassocyninae) from the Bolivian Altiplano. Journal of Paleontology 57, 1061-1066.

Wroe, S., McHenry, C. & Thomason, J. 2005. Bite club: comparative bite force in big biting mammals and the prediction of predatory behaviour in fossil taxa. Proceedings of the Royal Society B 272, 619-625.

Darren Naish is a science writer, technical editor and palaeozoologist (affiliated with the University of Southampton, UK). He mostly works on Cretaceous dinosaurs and pterosaurs but has an avid interest in all things tetrapod. His publications can be downloaded at darrennaish.wordpress.com. He has been blogging at Tetrapod Zoology since 2006. Check out the Tet Zoo podcast at tetzoo.com!

More by Darren Naish