Gallery and print store

Sunday, 9 October 2016

Exposed teeth in dinosaurs, sabre-tooths and everything else: thoughts for artists

Bear-sized gorgonopsid Inostrancevia latifrons. Sabre-teeth? What sabre teeth?
It is something of a trope that prehistoric animals must bare their teeth in palaeoart, even when their mouths are closed. Historically, the majority of palaeoartists covered the teeth of their subjects with lips, cheeks or other types of tissues and only select species – sabre-toothed carnivorans or mammoths – were depicted with exposed tusks or sabre-canines. This changed when artists working in the 1980s and 1990s - Paul, Hallett, Stout - and a certain 1993 movie started showing predatory dinosaurs with toothy overbites and perpetually exposed teeth. This convention has since expanded to all kinds of prehistoric animals, and some galleries of Deep Time now have more toothy grins than a holiday photo album. Theropod dinosaurs in particular are almost always shown with alligator-like overbites that perpetually expose their upper teeth, the large canines of stem-mammals protrude over their lower jaws, and even herbivorous animals with relatively unimpressive dentition (like sauropods) are shown without lips or other forms of dental covering.

Many words – mostly published at blogs, online mailing lists and social media - have been typed to discuss the credibility of lipless palaeoart, but the subject has traditionally received only cursory attention from academics. Happily for artists, this is starting to change. A small set of literature exists which debates the presence of extra-oral tissues in dinosaurs (e.g. Ford 1997; Knoll 2008; Morhardt 2009; Keilor 2013; Reisz and Larson 2016), and most this agrees that some sort of soft-tissue - at least 'lips' - covered their teeth. However, a running theme of these works is that reliably inferring soft-tissues of the face is not a simple task, and we really need more data to be sure of anything. Work on more recent fossil mammals shows more reliable inferences (e.g. Wall 1980; Antón et al. 1998), obviously benefiting from soft-tissue data from a range of extant, close relatives. New insights on the evolution of mammal cranial nerves are helping to understand the development of sensitive lips and cheeks in stem-mammals (Benoit et al. 2016). It's still early days for understanding fossil facial tissues, but at least it feels like we're off the line.

Collectively, there seems to be recognition among the academics interested in this topic that understanding the tooth coverage of fossil animals lies largely in understanding living animals. Attempts to understand tooth exposure from skulls alone - through making inferences about tooth size, jaw closure and speculations on how extensive soft-tissues can be before they become untenable - do not consider all necessary data. For example, Prehistoric Times palaeoart adviser Tracy Ford (1997) looked solely at the skulls of predatory dinosaurs to infer the absence of lips, suggesting their teeth were so long that they would pierce lip sheathing once the jaws were closed. This study assumed that predatory dinosaurs closed their mouths to the extent that the teeth of the lower jaw contacted the roof of the mouth, and that the preserved tooth configuration was the condition in life. These points are common issues raised against lipped dinosaurs, but there are several major problems. Dissections and CT scans of reptile heads show that jaw muscles and other soft-tissues have a major influence on mouth closure, to the extent that reptile jaw skeletons are typically loosely closed under their skin, even when the mouth is fully sealed. Taphonomic studies show that teeth slip readily from their sockets after death and often fossilise in far more vampiric states than they were in life. And undermining this further is that no extant taxa with lipped jaws were used to calibrate a limit for oral soft-tissues. Arguments about tooth coverage based on simply looking at skulls, without detailed consideration of modern animals and their anatomy, border on being arguments from incredulity : "I don't believe the anatomy could do that."

Modern animals and their tooth coverage

For an upcoming project, I've been trying to crystallise my approach to restoring ancient animal facial tissues, and deciding whether to cover their teeth or not is an important part of that discussion. I've been deliberately broad in this assessment to attempt to try to sort the wood from the trees: discussions of oral tissues can sometimes get lost in the minutiae of tissue types, uncertain osteological correlates and so on - and many of these discussions result in the same answer: they can't be resolved with current data. That's not to say they aren't important discussions, but it's helpful to step back to see if we can answer the simpler questions as well: what gauge of teeth can be covered by oral tissues? When are teeth actually exposed? And what questions should we, as palaeoartists, be looking to answer when restoring facial tissues?

Reviewing literature and galleries of modern animals, we can see that overwhelming majority of living tetrapods have covered teeth, including all amphibians, most mammals and most reptiles (excluding birds, naturally. Hey, if they wanted to be involved they shouldn't have lost their teeth). Exposed teeth are actually really rare, and a character completely absent in many major clades. The soft-tissues involved in covering the teeth are variable, but 'lips' – either slightly fleshy margins of skin, or skin overlying true muscle - are so universal among tetrapods, as well as living relatives like lungfish, that we might assume lip tissues of some kind were ancestral to the group, and breaching these with large teeth is a derived condition evolved independently in a minority of lineages. Crocodylians are the only living tetrapods with fully exposed teeth, but it's increasingly obvious that they're also pretty specialised/derived/downright weird (Grigg and Kirshner 2015). Far from being 'living fossils' frozen in evolution, they have so many anatomical nuances and specialisations that their use as model organisms for other extinct taxa is increasingly questionable. This applies to aspects of their facial anatomy too - we’ll discuss this in more detail below.
Fossil big-tooths - species almost universally depicted with exposed teeth - versus modern animals with huge, but completely covered teeth teeth. A, Inostrancevia latifrons; B, Tyrannosaurus rex; C, Smilodon fatalis; D, crocodile monitor Varanus salvadorii; E, mandrill Mandrillus sphinx; E, hippopotamus Hippopotamus amphibius. With the exception of Smilodon, the fossil taxa are out-toothed by the extant animals, and yet we know their oral tissues can accommodate their teeth without problem. Blue lines approximate lip margins in living species. A, after Kemp (2005); E, after Goldfinger (2004).
Looking inside animal heads (above) shows that facial soft-tissues can cover very, very large teeth – perhaps much larger than we might intuitively expect. Examples from a range of tetrapods – including rhinoceroses, sloths, tapirs, mandrills, baboons, camels, tuataras, snakes, peccaries, bullfrogs, hippopotamuses, monitor lizards, clouded leopards, numerous rodents and others – show that large fangs, robust tusks and other forms of enormous dentition can be retained within lips or cheeks. These large teeth are truly ‘hidden’ without bulges, changes in lip direction or other features to betray their presence, and are thus undetectable unless their owners open their mouths (and sometimes not even then). Many people are shocked by the size of animal teeth when they see their skulls, and the savagery of mammalian herbivore dentition – horses and camel fangs, rhino tusks, baboon canines - are particularly startling.

We owe many of these surprises to animal lips, which are generally much more extensive than we casually assume. Large teeth can slide into soft-tissue sheaths located between gums and lips, and these are quite visible in the open mouths of some species. Amphibians, lizards and many mammals have upper and lower lips of similar size which meet over the teeth and sheaths can form on either jaw, but some mammals – including most carnivorous forms - have very large, fleshy upper lips over thinner, tightly-bound soft-tissues of the lower jaw (Antón et al. 1998). In these species, the canine teeth overbite the lower lip but the upper ‘over-lip’ is large enough to obscure the fact that the tooth is outside the lower mouth tissues. I am unaware of a reversed situation with the lower lip covering a thin upper lip: this may reflect the fact that overbiting dentition is much more common than underbiting. Regardless of the specific configuration, it is clear that we should not underestimate the capacity for facial tissues to obscure even very large, sharp and ferocious-looking teeth. The assumption that all conspicuous teeth of fossil animals were on display in life is thus problematic and does not agree with what we can observe in modern animals (below).

Applying palaeoart-esque considerations of oral tissue capacity to modern mammals suggest hippos are giant hogbeasts and mandrills evolved in Mordor. Restoring modern animals using palaeoart approaches is a completely original concept which in no way owes anything to some book called All Yesterdays (Conway et al. 2013).
When do teeth breach the confines of soft-tissue? Mostly, it seems teeth used to process food remain covered. Mammal tusks and the exposed canines of certain deer are not directly involved in food processing, although this is not to say they are non-functional overall (e.g. elephants use their tusks to break branches, dig, topple trees; deer fight with their large canines). It seems that teeth of extreme size relative to the rest of the dentition are most likely to escape covering with soft-tissue, and it helps – though is not mandatory – if they grow obliquely or directly away from the jawline (this accounts for the majority of living mammal tusks). Teeth can remain covered even when their tips extend to the dorsal or ventral limits of the jaw skeleton, so long as they are aligned more or less vertically within the jaw (e.g. the mandrill skull illustrated above).

What's up with crocodylians?

The elephants – or rather large semi-aquatic reptiles – in the room here are crocodylians: why do they have exposed teeth when all other tetrapods have largely covered mouths? Their teeth are not overly large, nor acutely angled. Some (Reisz and Larson 2016) have argued crocodylian dentition is only possible because of their semi-aquatic habits. The (unpublished, currently conference abstract only) Reisz and Larson hypothesis is that exposed teeth – specifically their enamel component – are at risk from desiccation and breakage without constant hydration from saliva or environmental water (Reisz et al. 2016). This is an interesting idea which potentially gives artists a useful, practical guide to restoring prehistoric animals: anything living outside water with enamel-covered teeth must have covered them with soft-tissue. Despite its unpublished status, this idea has already chimed with some quarters of the online palaeoart community who're restoring anything with enamel-covered teeth with full sheathing.

We need to talk about enamel and exposed teeth. The exposed canines of male wild boars, Sus, have enamel (white shading) coatings on 3/4 of their surface, despite being exposed (dentine is dark grey, cementum is light grey). What does this mean for the enamel desiccation hypothesis outlined below? Image from Hillson (2005).
However, this proposal may not be as simple to implement as it first appears. For one thing, there is a real lack of consistency in tusk composition in living animals (see Hilson 2005). It is true that, as noted by Reisz et al. (2016), the tusks of elephants have caps of enamel and cementum that wear off rapidly, leaving their tusks composed of dentine alone. This would seem to support the desiccation hypothesis, it implying that enamel is a liability outside of the jaw soft-tissues. However, living elephants may not be typical in lacking enamel on their tusks, there being fossil and living mammals which do have substantial enamel components on their exposed teeth. For example, the tusks of several gomphothere species have broad bands of enamel along their lateral surfaces, even as adults (Padro and Alberdi 2008), while the canines of male musk deer are enamel covered on the external surface. The tusks of male wild boars and warthogs only bear dentine on the posterior surface and wear facet, the rest of these large, exposed teeth being covered in enamel. In all these species these are not just small bits of enamel that get worn off, but sustained coatings that persist on the tooth indefinitely and influence tooth wear (Koenigswald 2011). To confuse things further, walruses have dentine tusks like elephants, despite their aquatic habits seemingly precluding desiccation as a risk for their teeth, and the spiralling tusks of another marine mammal, the narwharls, are covered in enamel. If there is a relationship between enamel and tooth exposure, it is clearly a complicated one: the presence of absence of enamel in itself seems to have little bearing on tooth exposure in at least modern mammals. (Readers interested in tooth composition should check out the second edition of Samuel Hillson's Teeth (2005), for its extensive documentation and illustration of mammalian dentition).

Musk deer, Moschus, canines in lateral and medial view. Note the (white) enamel layer on the lateral surface, but dentine (grey) on the medial. From Hillson (2005 - the scale bar is likely erroneous!).
Our second reason to be sceptical of the enamel desiccation hypothesis concerns crocodylian behaviour. It is not widely appreciated that several crocodylians species ‘hibernate’, or more accurately aestivate, for months at a time in dry underground burrows during the hottest summer months (Grigg and Kirshner 2015). During these intervals they do not access water at all. Other, South American species spend dry spells as fully terrestrial carnivores, abandoning aquatic habits and obtaining water largely from the prey they kill (Grigg and Kirshner 2015). These states have to be explained against the suggested need to frequently moisten crocodylian teeth, because they suggest dental desiccation is not as risky as we're all assuming it is. Alternatively, it suggests that the requirement for hydration is so relaxed – literally months can pass without getting the teeth wet – that it probably has little influence on tooth anatomy.

Furthermore, there are important caveats about crocodylian facial tissues that we have to factor into any discussion of their lipless configuration. Crocodylian faces are far more specialised and unusual than they first appear, and this may factor into their lipless mouths. Their highly keratinous facial skin undergoes a developmental pathway unlike that of any other amniote (their facial skin is essentially have one, highly 'cracked' scale) (Milinkovitch et al. 2013) and their heads are riddled with hyper-sensitive Integumentary Sense Organs (ISOs). ISOs are another unique crocodylian feature and are attuned, among other things, to sensing tiny vibrations in water (Soares 2002, see Grigg and Kirshner 2015 for a recent overview). In at least some parts of the crocodylian skull ISOs are situated over tiny foramina, presumably housing nervous tissues, and the overlying epidermis is thinned, with a reduced keratin component, to enhance their sensitivity (Soares 2002). We can thus see that ISOs do have a role to play in configuring crocodylian skin, and they present many questions that palaeoartists should be interested in. Are ISOs an important reason for crocodylian faces having such tight, contour hugging and lipless skin? Do the major functional and developmental distinctions of croc faces explain the lack of crocodylian lips? It might explain why virtually no other aquatic tetrapods have abandoned lips - aside from the the odd (and perhaps only) exception like the South Asian river dolphin*, there are no whales, snakes, seals or otters with crocodylian-like, fully exposed teeth. And given that no other lineages have osteological correlates for ISOs, should we put huge caveats around using crocodylians as models for facial tissues in anything other than their own ancestors? I don't know if anyone has answers to these questions yet, but they're food for thought when using crocodylians as ammunition for lipless reconstructions of fossil animals.

*Thanks to Ádám Lakatos for pointing out the toothiness of some river dolphins!

It's still very early days for the enamel/oral covering hypothesis, but modern animals suggest that interpretations of enamel precluding extraoral teeth are definitely more complicated than they first appear, and may even be flawed. If so, the simple presence of enamel on the teeth of fossil organisms may not be as useful to artists as some are currently suggesting. But this conclusion is preliminary, and we need to wait for this idea to mature before it's shot down entirely. We know, for instance, that there's more than one type of enamel among vertebrates. Reptilian enamel, for instance, is both thinner and of different microstructure to mammalian enamel, and these clades have rather different approaches to tooth longevity. This may mean something for enamel desiccation and long-term tooth exposure, and we may think differently on this matter once this research has been completed.

Predicting tooth exposure in fossil species

Fully 'lipped' gorgonopsids and theropods: maybe not be as exciting as their toothy variants, but are they more credible? Well... if modern animals are anything to go by, probably.
All this said, what can we say about the decisions to show prehistoric animals with exposed teeth? My reading of modern tetrapods is that covered teeth is their ‘default’ configuration, and we should apply the same logic to extinct animals. If so, maybe only the more extreme examples of fossil dentition should qualify for perpetual display. Perhaps instead of asking ‘does this animal have lips?’ we should ask why they should not have them. We have to concede that the dentitions of many fossil animals frequently shown with exposed teeth – particularly theropod dinosaurs, gorgonopsids and other carnivorous stem-mammals – are relatively no larger, and in some cases a great deal smaller, than those enclosed inside the oral tissues of living animals, especially once taphonomic tooth slippage is corrected (above). For these species, it is very difficult to justify why their teeth should not be covered.

If this is so, only especially long teeth which project a considerable distance from the margins of the skull and lower jaw should be considered strong candidates for permanent exposure. Select examples might include the canines of certain mammalian carnivores (e.g. Smilodon and other machairodont felids), the tusks of fossil elephants and their relatives, and the larger tusks of dicynodonts. We should also note those fossil reptiles – such as certain crocodyliformes, pterosaurs and marine reptiles – where entire toothrows are composed of dentition so long that their tips extend well beyond the margins of the jaw skeleton. Such extensive dental apparatus would seem to preclude the development of any sheathing tissues, at least akin to those exhibited by from modern animals, and these animals probably had fully exposed toothrows in life. Of course, this conflicts with the observation that food-processing teeth are almost always covered in the modern day. However, we can defend this approach by arguing that their morphology gives a strong reason for ignoring this guideline: it answers the "why we shouldn't give them lips?" question.

The large, procumbent dentition of plesiosaurs and certain pterosaurs argues against them being sheathed in life, although I do wonder if some plesiosaurs are in a 'grey area' here. Could animals like Leptocleidus (right) have covered its teeth with lip-like tissues? Hmm....
We might also set aside this guideline when extant relatives of modern forms provide us with means to predict unusual lip anatomy. For instance, the aforementioned ‘over-lip’ of modern mammalian carnivores is common enough across this group to assume it was present in their fossil relatives. Because we understand how the lips of these animals work, we can make more specific predictions concerning tooth exposure in species with particularly impressive teeth. Thus, we can look at classic reconstructions of machairodontid cats like Smilodon with perpetually bared fangs as reasonable because, unless their lips were arranged differently to virtually all their living relatives, that’s simply how their lip tissues would respond to a massive set of canines. And yes, I'm aware of Dunae Nash's recent discussions about sheathing Smilodon: given that this rests heavily on enamel being a no-no in exposed teeth, I'm unconvinced for the reasons explored above.

The concluding caveat

Of course, it must be reinforced that these are just guidelines - and guidelines based on my own qualitative studies, nonetheless, your mileage with them may vary - and there are exceptions to the suggestions made above. As is well known, for all the suggestion that restoring sabre-toothed cats with exposed teeth is reasonable, one living cat species – the clouded leopard – does cover a set of long upper canines in a lower lip sheath. We would not predict this based on other cat species and, if known only from fossils, clouded leopards would probably be restored with a slightly exposed canine. Likewise, the exposed tusks of some deer are not especially massive, and if we followed the suggestions above we'd probably cover them up in a reconstruction. But palaeoart is ultimately a game of prediction and probability, attempting to restore what is most likely to fill gaps in our data, and any game of odds will have some failures. That’s not to say we shouldn’t ignore these exceptional examples - they show that guidelines can't be trusted all the time - but it makes sense for us to know where the guidelines are in the first place. As with all aspects of palaeorestoration, all of us stand a chance to be proved wrong about our artistic decisions: if and when that happens, the best we can hope for is to have been wrong for the right reasons.

This blog post was covered by Patreon

The paintings and words featured here are sponsored by a group of animals which also have sheathed dentition, Patreon backers. Supporting my blog from $1 a month helps me produce researched and detailed articles with paintings to accompany them, and in return you get access to bonus blog content: additional commentary, in-progress sneak-previews of paintings, high-resolution artwork, and even free prints. Sign up to Patreon to get access to this and the rest of my exclusive content!


  • Antón, M., Garcia‐Perea, R, & Turner, A. (1998). Reconstructed facial appearance of the sabretoothed felid Smilodon. Zoological Journal of the Linnean Society, 124(4), 369-386.
  • Benoit, J., Manger, P. R., & Rubidge, B. S. (2016). Palaeoneurological clues to the evolution of defining mammalian soft tissue traits. Scientific reports, 6.
  • Conway, J., Kosemen, C. M., Naish, D., & Hartman, S. (2013). All yesterdays: unique and speculative views of dinosaurs and other prehistoric animals. Irregular Books.
  • Ford, T. L. (1997). How to Draw Dinosaurs. Give Theropods no Lip! Prehistoric Times, 25, 49-50.
  • Hillson, S. (2005). Teeth, Cambridge Manuals in Archaeology Series. Cambridge University Press, Cambridge, 373.
  • Goldfinger, E. (2004). Animal anatomy for artists: The elements of form. OUP USA.
  • Grigg, G., & Kirshner, D. (2015). Biology and evolution of crocodylians. Csiro Publishing.
  • Keillor, T. (2013). June, in the Flesh: The State of Life-Reconstruction in Paleoart. In: Parrish, J. M., Molnar, R. E., Currie, P. J., & Koppelhus, E. B. (eds). Tyrannosaurid Paleobiology, Indiana University Press. 157-176.
  • Kemp, T. S. (2005). The origin and evolution of mammals. Oxford University Press.
  • Koenigswald, W. V. (2011). Diversity of hypsodont teeth in mammalian dentitions—construction and classification. Palaeontogr. Abt. A, 294, 63-94.
  • Knoll, F. (2008). Buccal soft anatomy in Lesothosaurus (Dinosauria: Ornithischia). Neues Jahrbuch für Geologie und Paläontologie-Abhandlungen, 248(3), 355-364.
  • Morhardt, A. C. (2009). Dinosaur smiles: Do the texture and morphology of the premaxilla, maxilla, and dentary bones of sauropsids provide osteological correlates for inferring extra-oral structures reliably in dinosaurs? (Doctoral dissertation, Western Illinois University).
  • Prado, J. L., & Alberdi, M. T. (2008). A cladistic analysis among trilophodont gomphotheres (Mammalia, Proboscidea) with special attention to the South American genera. Palaeontology, 51(4), 903-915.
  • Reisz, R. R. & Larson, D. (2016) Dental anatomy and skull length to tooth size rations support the hypothesis that theropod dinosaurs had lips. 2016 Canadian Society of Vertebrate Paleontology Conference Abstracts, 64-65.
  • Soares, D. (2002). Neurology: an ancient sensory organ in crocodilians. Nature, 417(6886), 241-242.
  • Wall, W. P. (1980). Cranial evidence for a proboscis in Cadurcodon and a review of snout structure in the family Amynodontidae (Perissodactyla, Rhinocerotoidea). Journal of Paleontology, 54, 968-977.

Monday, 12 September 2016

A salute to the Erythrosuchidae

Two Garjainia madiba decide who gets the table scraps. The reconstructions here are modified from the life reconstruction I provided for Gower et al. (2014).
I find erythrosuchids, large, big-headed Triassic archosauriforms, very charismatic fossil animals. If nothing else, it's hard not to admire their no-nonsense approach to carnivory. Take a fairly standard reptilian bauplan, weld an oversize theropod dinosaur face to the front, then point it at the things you want to die: simple. They're the Mesozoic equivalent of mounting a howitzer on a golf cart and calling it a tank. We might question the rudimentary nature of the design, but we can't argue with the results.

Alas, erythrosuchids don't get as much love from palaeoartists or outreach projects as they deserve. Their marriage of a proportionally huge, sharp-toothed skull with a crocodile- or lizard-like body is unlike anything around today and it's difficult not to wonder how they functioned as living animals. Closer inspection of their anatomy reveals more sophistication than we might assume from the few illustrations of these animals available online or in books, and it seems that their role in Mesozoic ecosystems and reptile evolution was an important one. These were a successful, abundant group of predators with an evolutionary run spanning the Early and Middle Triassic (12 million years in total) and a near cosmopolitan distribution. Moreover, they remain important species for understanding the early evolution of archosaur-line reptiles. They really do have a lot going for them, but they just haven't quite caught public imagination.

A few years ago I was commissioned to reconstruct the small(ish), early erythrosuchid Garjania madiba for David Gower and his colleagues for their 2014 descriptive paper (below). The brief was for a straight illustration of the animal rather than a restored scene, and I promised the team I would put this reconstruction in a landscape one day. Two years later, I've finally got around to it: the results are above. Posting this painting seems like as good an excuse as any to lavish some much needed attention on these most encephalised of reptiles, so let's get stuck in.

G. madiba reconstruction from Gower et al. (2014). Note prominent bosses on the face, a characteristic feature of this species.

What, exactly, is a erythrosuchid?

You can find erythrosuchids in Triassic rocks on every continent except North America and Antarctica and, although relatively complete specimens are not common, many species are represented by large inventories of bones. Despite this relative glut of material, the classification of erythrosuchids - from the fine anatomical characteristics of the group, to their position in the reptilian tree and the number of species contained in the clade - has been the subject of long-standing, ongoing discussions among palaeontologists. Older erythrosuchid literature is confused by a multitude of different classifications which entwine erythrosuchids with other large-headed, carnivorous archosauriforms such as raisuchians and proterosuchids. Researchers have long realised the problems with these schemes, but unpicking the relationships of these groups and other early archosaur-line reptiles has been tricky. With the arrival of extremely detailed and well sampled cladistic analyses of archosauromorphs (e.g. Nesbitt 2011; Ezcurra 2016) we might be moving towards greater consensus on the systematics of these animals, however. In modern schemes, erythrosuchids are recovered as non-archosaur archosauriforms close(ish) to the base of Archosauria. More specifically, they are the sister clade to the the Eucrocopoda, the large clade that contains the likes of Euparkeria and proterochampsids, as well as the true archosaurs (Ezcurra 2016).
Erythrosuchus africanus skull, restored by Gower (2003). Note the extremely robust construction of the bones and expanded areas for neck muscle attachment.
Several erythrosuchid species are well known: Erythrosuchus africanus from the Middle Triassic of South Africa, Garjainia prima from the Early Triassic of Russia, and Shansisuchus shansisuchus (that's not a typo) from the Middle Triassic of China. These species are represented by associated remains as well as large numbers of fragmentary referred specimens, and allow for a relatively complete insight into their overall form. The largest taxa, like Erythrosuchus, are big animals with head-tail lengths approaching 5 m - the length of a good-sized car - and even small taxa like Garjainia are over 2 m long. The most arresting aspect of eyrthrosuchid anatomy is, of course, their skulls (above). Superficially theropod-like, these long, deep and robust structures are sub-rectangular in lateral view, but taper markedly towards the snout in dorsal or ventral aspect. These animals are yet another reminder that restoring fossil animals needs more than a lateral view of a skeleton: those massive skulls are considerably narrower than we might expect. Their teeth are thecodont, large, serrated and recurved. A characteristic of the group is the complicated shape of the upper jaw, where the jaw tip is vertically displaced from a ventrally bowing maxillary region (Parrish 1992), creating something of a 'notch' towards the front of the jaw. Beneath this, the mandible has a slightly dorsoventrally expanded tip, as well as a swollen posterior region. At least the skull of Erythrosuchus is essentially akinetic, although minor movements of some bones may have been possible (Gower 2003). Although erythrosuchid skulls are fairly conservative in morphology, some species were not above frivolous accessorising: prominent bosses above and below the eye are known from Garjainia madiba (Gower et al. 2014 - see reconstructions, above), and Pickford (1995) reports a long, low boss on the snout of an undescribed Karoo Basin specimen.

Although erythrosuchid skulls were almost certainly pneumatised in some areas, the largest opening in the skull is not, as we might expect in such large headed animals, anything to do with a pneumatic cavity. Rather, it's the lower temporal fenestra, an opening typically associated with allowing bulges of the jaw adductor muscles. This, as well as the presence of a small sagittal crest between the superior temporal openings (which overly the same muscle block) and the depth of the posterior mandible likely betrays the presence of massive adductor muscles in temporal region of the skull. Eryhtrosuchid skull bones certainly look sufficiently robust to withstand powerful biting, the bones forming the temporal fenestra, jaw and orbital margins being extremely massive and thick and tightly interlocking with complex sutures between each bone. Interestingly, Shansisuchus has the same partly invaded orbit shape that Henderson (2003) linked with reinforcement against heavy bite forces in theropod dinosaurs: perhaps similar buttressing was taking place in these Triassic reptiles

The dorsal extent of the occipital face in Eryhtrosuchus africanus, posterior view. The rounded flanges at the top poke above the rest of the skull, and perhaps indicate expanded neck muscles in this and other species. From Gower (2003).
The posterior surface of the skull is interesting. Rather than the relatively flat surface we see in most animals, the posterior erythrosuchid skull is recessed so that several aspects of the skull - the jaws and lateral extents of the occipital surface - extend further back than the vertebral/skull joint. The area which anchored the neck musculature extended across this recessed surface, even exceeding the dorsal margins somewhat by means of a pair of semiscircular flanges projecting above the rest of the skull (visible in at least Erythrosuchus and Garjainia - see above). Assuming a typically reptilian muscle plan, these indicate that muscles anchoring above the skull-neck articulation were larger than usual, as might be expected for animals with ginormous heads. Similar dorsal expansion of the occipital region is seen in tyrannosaurids, and is also thought to reflect large cervical musculature (Paul 1988). It thus seems the vertebrae and posterior skull of erythrosuchids were deeply buried in neck tissues, befitting animals with a giant head to support and utilise in predatory acts. But I wonder if all this support and strength compromised the mobility of the skull-neck joint somewhat. Moving the neck articulation forward to sit within the boundaries of the skull likely shortened the length of the skull flexor muscles, as well as buried the joint in masses of potentially restrictive muscle and bone. Motion of the head may have been limited at the front of the neck, then, but unfortunately for erythrosuchid prey, the size of the shoulder skeleton and stoutly built humeri suggest this was accounted for with powerful muscles at the base of the neck, as well as forelimbs able to shove the forequarters around at speed. Dashing left or right against a charging erythrosuchid was unlikely to save you from a nasty, gigantic and powerful bite.

Behind the skull we see a fairly typical Triassic archosauriform body (below). The neck is short, and especially so in some of the larger species, and the majority of the vertebrae are adorned with tall neural spines: these almost certainly provided anchorage for axial musculature related to supporting the head and back. The pectoral elements, which are also employed somewhat in neck musculature, are also robust. Their tails are moderately long, with deep chevrons in the anterior region likely related to hindlimb musculature. Behind these, the tail becomes rather slender. Gower (2001) proposed that Erythrosuchus vertebrae possessed pits and depressions possibly related to the development of post-cranial pneumaticity, the first found outside of pterosaurs and dinosaurs. This would be a significant find, telling us something of erythrosuchid lung structure as well as the early evolution of postcranial pneumaticity in archosaur-line reptiles. However, both O'Connor (2006) and Butler et al. (2012) argued against this interpretation, noting that the features in question were not associated with internal cavities, thus failing to meet criteria for structures of pneumatic origin. An important caveat to this, however, was raised by Butler et al. (2012): the phenomenon of pneumatic tissues invading vertebrae and other postcranial bones almost certainly did not evolve in one swoop. Its earliest stages may have simply been pneumatic tissues 'pushing' against external bone walls, forming pits and cavities, rather than invading them entirely. If so, the sort of thing Gower (2001) found in Erythrosuchus might be what we'd expect of early stage, postcranial pneumaticity. So while we have to concede that these structures do not meet our current definition of a postcranial pneumatic structure, perhaps we also need to learn more about the early evolution of postcranial pneumaticity before this hypothesis can be ruled out entirely.

Mounted Garjainia prima skeleton as mounted at the Paleontological Institute, Moscow. Certain aspects of this skeleton are reconstructed or sculpted, so take some details with a pinch of salt. From Ivakhnenko and Kurochkin (2008).
The limbs of erythrosuchids are not, to my knowledge, completely known from any species but their major limb bones are powerfully built and surprisingly lengthy: you could never call them 'long-limbed', but they are not the stumpy-legged animals we often see them reconstructed as. Their hands and feet are poorly known. Rare examples of erythrosuchid ankles are thought to indicate an mesotarsal condition (Gower 1996), and their pelves show signs of advanced features that we see developed further in true archosaurs. These features led to our G. madiba reconstruction having semi-erect hindlimbs, while the forelimbs remained sprawling. The typical pose of erythrosuchids remains to be determined from further study of their limb bones.

A point of contention among researchers is whether or not erythrosuchids had osteoderms. Two examples of such structures have been found in association with a specimen of Erythrosuchus, but they show no consistency in their morphology (Gower 2003). Moreover, the extensive inventory of Erythrosuchus and other erythrosuchids have yet to show additional evidence of dermal bones (Ezcurra et al. 2013). The safe bet, for the time being at least, is to assume these reptiles did not have osteoderms, and that those previously referred to the group were a fluke association from another animal.

The life and times of Triassic big-heads

We have much to learn about many aspects of erythrosuchid palaeobiology: details of their dietary preferences, locomotor mechanics and likely habitats remain only provisionally researched. Much of what we've learned about their lifestyles comes from 'bigger picture' assessments of Triassic diversity and faunal turnover, so we can only paint a broad-brush picture of their ecology at this time. That's not to say we have no specific palaeobiological insights into these animals, however. For instance, there is consistent histological evidence that erythrosuchids grew quickly, perhaps at rates comparable to pterosaurs and dinosaurs, until they reached reproductive maturity (de Ricqlès et al. 2008; Botha-Brink and Smith 2011; Ezcurra et al. 2013). Given that this trait is not limited to erythrosuchids among Early and Middle Triassic reptiles, this is one reason it's thought that archosaur-line reptiles may not be ancestrally ectothermic. Whatever the cause, rapid growth may have played some role in the success of erythrosuchids and other reptiles as ecosystems were rebuilt in the early Mesozoic (Sookias et al. 2012).

Erythrosuchid ecology remains only lightly investigated, but they have been considered arch terrestrial predators by some (Sennikov 1996 - see below). Interestingly, their size puts them among the largest terrestrial animals known from their respective faunas (Sookias et al. 2012). This is unusual: in post-Middle Triassic ecosystems we generally find herbivores are the largest animals in terrestrial ecosystems, so what's going on here? It's thought that physiological distinctions between large Early-Middle Triassic reptiles and the synapsid herbivores they coexisted with may explain the size difference (briefly summarised, archosauriform growth rates and respiratory anatomy may have permitted larger overall body size than therapsids - see Sookias et al. 2012), but how did this translate into ecological balance? Energy is lost as it is transferred between species in food webs, so how did populations of relatively 'giant' top-tier erythrosuchids sustain themselves on consistently smaller prey? Perhaps they were simply comparatively rare, or very energy-efficient, or maybe they supplemented their diet with non-terrestrial food items - did they also take food from aquatic realms, perhaps?

An Early Triassic terrestrial food web, reconstructed for the Yarenga Formation by Sennikov (1996). In this scheme, most things ended up in the bellies of erythrosuchids or rausuchians.
Speaking of aquatic habitats, the concept of erythrosuchids as strictly terrestrial predators is not the only interpretation of their habits. Indeed, for much of the 20th century erythrosuchid proportions were considered evidence of aquatic or semi-aquatic habits: their huge heads and robust limbs were thought to permit only cumbersome, laboured movement on land (see Ezcurra et al. 2013 for a brief review). The words offered by Reig (1970) paint an excellent summary of these older interpretations: "We doubt that bulky and clumsy animals like Erythrosuchus and Shansisuchus should be considered very active animals... It is more likely that they were inhabitants of swamp marshes, able to prey upon big, slow herbivorous vertebrates, inhabiting the same environments, which could be caught by a relatively slow and heavily built predator" (p. 261). Potentially further evidence of semi-aquatic lifestyles are the relatively thick limb bone walls common to all erythrosuchids, these being comparable in thickness to those of modern alligators (Botha-Brink and Smith 2011; Gower et al. 2014).

In recent years, however, erythrosuchids seem to have been perceived as more terrestrial animals (Sennikov 1996; Botha-Brink and Smith 2011; Ezcurra et al. 2013). Their thick bone walls are explained as being a consequence of their large size rather than aquatic habits (Botha-Brink and Smith 2011) and the deficit of obvious aquatic adaptations in their skeletons has been noted by several authors (Botha-Brink and Smith 2011; Ezcurra et al. 2013; Gower et al. 2014).

Aquatic, semi-aquatic or fully terrestrial? This guy's meant to have taken a dip in the water, but was it intentional or accident? We may not have the data to say exactly what erythrosuchids did for a living yet.
All this said, I must admit to desiring more work in this area. The habits of strange Triassic animals are difficult to fathom in many instances, and we're yet to see particularly comprehensive assessments of the most basic elements of erythrosuchid functional anatomy, let alone application of modern techniques like isotope analysis, stress modelling of jaws and so on to this problem. My gut feeling - and thus in no means a basis for a hypothesis - is open to both interpretations of erythrosuchid habits, and I wouldn't be surprised if terrestrial and aquatic prey were on their radars. I'm suspicious about the weight of the head being a problem for terrestrial locomotion. A decade of looking at terrestrially-competent, large-headed pterodactyloid pterosaurs and recent monkeying about with mass fractions of giant-necked Tanystropheus suggest our intuitive grasp of front-heaviness might be poorly calibrated. Animal heads and necks are often much lighter than we think in contrast to torso and limb masses, and we should remind ourselves that erythrosuchid skulls are actually quite narrow, presumably well-pneumatised structures. This is the sort of thing that can be relatively easily investigated using digital models, and we might hope this approach is applied to erythrosuchids in future. But if that supports a terrestrial habit, the notched upper jaw and swollen mandibular tip of erythrosuchids argues contrarily: similar jaw tips are seen in fish-eating animals like modern crocodylians and pike conger eels, as well extinct presumed fishers such as spinosaurids and some pterosaurs. Might this not imply that small swimming animals were sometimes eaten by erythrosuchids, too? Lest we forget, animals do not necessarily need to be dedicated swimmers to be able to eat aquatic prey. There's a lot of scope for further work and investigation here, and it would be great to see some dedicated functional assessments and ecological investigations of erythrosuchids in future.

I love it when a bauplan comes together

Perhaps one of the most interesting things mentioned recently about erythrosuchids is how little their postcrania differs from those of other archosauriforms, despite their substantial cranial modifications (Ezcurra 2016). This is something we see again and again in Triassic reptiles: relatively conservative bodies with highly localised outlandish anatomy, and is true even for the weirdest Triassic creatures. For example, Tanystropheus isn't that strange aside from its incredible neck, and (what we know of) the body of Sharovipteryx is not that atypical in spite of its leg-wings. I wonder if Triassic animals get the short shrift in popular circles because they're viewed as boring 'also rans' taxa which evolved strange, untenable anatomies but without moving too far from a typically 'reptilian' visage.

But perhaps what we're seeing with these animals is far more interesting than it first appears: a display of the intrinsic adaptability of the archosauromorph bauplan, and how applicable it was to many lifestyles with only localised modification. We can be particularly impressed with erythrosuchids because of their rapid evolution so early in the Triassic: they very quickly and successfully jumped into the niche of large, hypercarnivorous apex-predator after the end-Permian extinction event, and then held that niche worldwide for 12 million years. The fact they did so without much additional modification to the postcrania is evidence that their success was not a fluke, and that the basal archosaur-line body plan was a strong one. Perhaps instead of looking at erythrosuchids and other Triassic archosauromorphs as those strange, but ultimately dull animals that struck it lucky before the more successful ones took over, we might view them as some of the earliest evidence that the archosaur-line bauplan had real potential, and a sign of what was to come.

Big blog posts about big headed reptiles need big support - thank goodness for Patreon

The paintings and words featured here are sponsored by a group of tetrapods with more modestly proportioned skulls, my Patreon backers. Supporting my blog from $1 a month helps me produce researched and detailed articles with paintings to accompany them, and in return you get access to bonus blog content: additional commentary, in-progress sneak-previews of paintings, high-resolution artwork, and even free prints. For this post, we'll be looking at the history of the painting up top, documenting its long 2-year journey from illustration to, er, a more detailed illustration. I'll also share the bizarre, sausage piglet monster version of Garjainia that you were never meant to see. Sign up to Patreon to get access to this and the rest of my exclusive content!


  • Botha-Brink, J., & Smith, R. M. (2011). Osteohistology of the Triassic archosauromorphs Prolacerta, Proterosuchus, Euparkeria, and Erythrosuchus from the Karoo Basin of South Africa. Journal of Vertebrate Paleontology, 31(6), 1238-1254.
  • Butler, R. J., Barrett, P. M., & Gower, D. J. (2012). Reassessment of the evidence for postcranial skeletal pneumaticity in Triassic archosaurs, and the early evolution of the avian respiratory system. PloS one, 7(3), e34094.
  • de Ricqlès, A., Padian, K., Knoll, F., & Horner, J. R. (2008). On the origin of high growth rates in archosaurs and their ancient relatives: Complementary histological studies on Triassic archosauriforms and the problem of a “phylogenetic signal” in bone histology. In Annales de paleontologie (Vol. 2, No. 94, pp. 57-76).
  • Ezcurra, M. D., Butler, R. J., & Gower, D. J. (2013). ‘Proterosuchia’: the origin and early history of Archosauriformes. Geological Society, London, Special Publications, 379(1), 9-33.
  • Ezcurra, M. D. (2016). The phylogenetic relationships of basal archosauromorphs, with an emphasis on the systematics of proterosuchian archosauriforms. PeerJ, 4, e1778.
  • Gower, D. J. (1996). The tarsus of erythrosuchid archosaurs, and implications for early diapsid phylogeny. Zoological Journal of the Linnean Society, 116(4), 347-375.
  • Gower, D. J. (2001). Possible postcranial pneumaticity in the last common ancestor of birds and crocodilians: evidence from Erythrosuchus and other Mesozoic archosaurs. Naturwissenschaften, 88(3), 119-122.
  • Gower, D. J. 2003, Osteology of the early archosaurian reptile Erythrosuchus africanus, Broom. Annals of the South African Museum, 110(1), 1 - 84.
  • Gower, D. J., Hancox, P. J., Botha-Brink, J., Sennikov, A. G., & Butler, R. J. (2014). A new species of Garjainia Ochev, 1958 (Diapsida: Archosauriformes: Erythrosuchidae) from the Early Triassic of South Africa. PloS one, 9(11), e111154.
  • Henderson, D. M. (2003). The eyes have it: the sizes, shapes, and orientations of theropod orbits as indicators of skull strength and bite force. Journal of Vertebrate Paleontology, 22(4), 766-778.
  • Ivakhnenko, M. F. & Kurochkin, E. N. (eds.) 2008. Fossil Vertebrates of Russia and adjacent countries. Fossil reptiles and birds. Part 1: A. Reference book for paleontologists, biologists and geologists. GEOS, 2008, 348 pp.
  • Nesbitt, S. J. (2011). The Early Evolution of Archosaurs: Relationships and the Origin of Major Clades. Bulletin of the American Museum of Natural History, 1-292.
  • O'Connor, P. M. (2006). Postcranial pneumaticity: An evaluation of soft‐tissue influences on the postcranial skeleton and the reconstruction of pulmonary anatomy in archosaurs. Journal of Morphology, 267(10), 1199-1226.
  • Parrish, J. M. (1992). Phylogeny of the Erythrosuchidae (Reptilia: Archosauriformes). Journal of Vertebrate Paleontology, 12(1), 93-102.
  • Paul, G. S. (1988). Predatory dinosaurs of the world: a complete illustrated guide. Simon & Schuster.
  • Pickford, M. (1995). Karoo Supergroup palaeontology of Namibia and brief description of a thecodont from Omingonde. Palaeontologia Africana, 32, 51-66
  • Sennikov, A. G. (1996). Evolution of the Permian and Triassic tetrapod communities of Eastern Europe. Palaeogeography, Palaeoclimatology, Palaeoecology, 120(3), 331-351.
  • Reig, O. A. (1970). The Proterosuchia and the early evolution of the archosaurs; an essay about the origin of a major taxon. Bulletin of the Museum of Comparative Zoology, 139(5), 229-292.

Wednesday, 31 August 2016

New paper: at last, a small pterosaur species from the latest Cretaceous

As the Cretaceous fossil record enters its final two stages - the Campanian and Maastrichtian - several unusual things seem to happen in the world of flying reptiles. Firstly, we see the end result of a steady drop off in pterosaur diversity leaving only azhdarchids - those often long-necked, long faced animals that we cover here with some regularity - with a strong, widespread fossil record. It's known that nyctosaurids and (thanks to recent discoveries) perhaps pteranodontids survived until the very end of the Mesozoic in at least two locations, but azhdarchids are globally distributed and dominate the pterosaur fossil record at this time. The overwhelming precedence of azhdarchids in the Late Cretaceous is an anomaly: at no other point in the pterosaur fossil record does one clade feature so prominently.

Secondly, Campanian and Maastrichtian pterosaurs are, without exception, pretty big animals. Many species from this time are renowned for their gigantic size: it's these stages which give us the famous 10 m wingspan, 250 kg colossi like Quetzalcoatlus, Arambourgiania and Hatzegopteryx, as well as a number of other giant azhdarchids which are too poorly known for generic titles. Coinciding with the evolution of the giants is a loss of small pterosaur taxa - those animals less than 2.5 m across the wings which are present, more-or-less, throughout the rest of pterosaur history. This shift in body size is sometimes interpreted as pterosaurs demonstrating 'Cope's Rule', the somewhat controversial proposal that species evolve towards large body size over time (Hone and Benton 2007; Benson et al. 2014). It's argued by some that competition from birds may be the driver behind this trend, as early avians fought small flying reptiles for ecological space and ultimately forced pterosaurs into larger sizes (e.g. Benson et al. 2014). Note that this concept is not without its detractors, including myself - I won't go into my reasons now but I plan to outline them here eventually.

Whether you agree with the bird-pterosaur competitive displacement hypothesis or not, we can't disagree that the end of the Cretaceous is almost entirely devoid of small pterosaur remains. Only a handful of specimens record small pterosaurs in the Campanian and Maastrichtian, and they're all tricky to work with. Aside from being highly fragmentary, some are controversially identified (such as Piksi barbarulna, an alleged small pterosaur from the Two Medicine Formation - see Agnolin and Varricho 2012 for the pro-pterosaur case) and others represent probable juvenile individuals (Godfrey and Currie 2005). Whatever it signifies, the lack of diminutive pterosaur specimens from the close of the Mesozoic is a real phenomenon of our fossil record, and any new specimen of a small, latest Cretaceous flying reptile has to be something to get excited about.

Enter: a new small, latest Cretaceous pterosaur specimen to get excited about

Title slide of my SVPCA 2016 talk, discussing the findings of Martin-Silverstone et al. 2016, out today. If you don't get the reference, you clearly get out too much, have too many friends and aren't watching enough crap TV.
It's this point where a new paper, published today by Liz Martin-Silverstone, myself, Victoria Arbour and Phil Currie comes in. Our new work, which you can check out without restriction at the open access journal Royal Society Open Science, presents a new small pterosaur fossil from the Campanian Northumberland Formation of British Columbia. The specimen number - RBCM.EH.2009.019.0001 - is pretty unwieldy, so I've been calling it the 'Hornby azhdarchoid' or the 'Hornby pterosaur' after it's discovery on Hornby Island, just off the coast of Vancouver. As you can see  below, the Hornby specimen is not pretty. Following our presentation of the fossil at SVPCA 2016, pterosaur guru David Unwin suggested we might have the ugliest pterosaur fossil on record (or at least tied the game). But while not well preserved, we do at least have several bones to play with: most of a humerus, three fused vertebrae (from the notarium, a set of fused shoulder vertebral elements), a few loose dorsal vertebrae and some other odds and ends that defy identification. This makes it the first set of associated bones of a small latest Cretaceous pterosaur, which is at least a step in the right direction for their paltry fossil record. For reasons discussed in the paper (concerning taphonomy, element size and likely identifications) we assume these remains represent one individual.

RBCM.EH.2009.019.0001, a fragmentary azhdarchoid pterosaur from the Campanian Northumberland Formation, British Columbia. It's, er, not the prettiest pterosaur specimen you'll ever see. Combination of figures from Martin-Silverstone et al. 2016.
I don't want to rehash the full gory details of our study here - please read the paper for the technical aspects - but instead want to outline our main points. The first thing to clear is that we've been careful to rule out an avian ID for the specimen. The Northumberland Formation contains several bird fossils and the quality of the specimen means that many obvious pterosaur features are missing. The Royal British Columbia Museum was kind enough to ship the specimen all the way from Vancouver, Canada to Southampton, UK just so Liz and I make a thorough assessment on this issue. Happily, we found the specimen to be very pterosaur like in every aspect (even as fragments, pterosaur bones are quite distinctive) as well as differing from Mesozoic birds in several ways. It particularly contrasts in having a notarium, which seem absent from Mesozoic birds (note that we compared the notarium element compared carefully with Mesozoic bird synsacra to be sure of our identification), as well as having a pterosaur-like, rather than avian, proximal humerus morphology. But we're not bird workers so, to be extra sure, we showed the material to fossil bird experts in Canada and the UK (including people who've identified and published on the Northumberland Formation avians). No-one we spoke to suggested an avian ID and, moreover, we are aware that other people with expertise in both birds and pterosaurs (including our paper editor) have seen the material and prefer a pterosaur ID. Based on our research and the testimonials of others, we're as confident as we can be that the Hornby fragments represent a pterosaur, not a bird.

We've identified the Hornby specimen as an azhdarchoid, and noted several features indicative of, but not conclusive to, an azhdarchid ID. We suspect the specimen is an azhdarchid because of its provenance and its basic anatomical characteristics, but the specimen does not contain the right bits to confirm an azhdarchid identity. Nonetheless, narrowing the specimen down to Azhdarchoidea allows us to estimate its body proportions and confirm that the specimen was indeed a small animal when it died. We estimated its wingspan using two methods factoring both the humerus and vertebrae, and each pointed to a wingspan between 1.4 and 1.6 m. That puts our pterosaur at a comparable size to a good sized-seagull and, while these are respectably-sized modern birds, this is small for a latest Cretaceous pterosaur. Rather than poking giraffes in the face, our little chap would only just be beyond predation risk from an average housecat (below). The only contemporary pterosaur competing with the Hornby azhdarchoid for size is Piksi, a poorly known possible pterosaur from the western US. Our new study lists a number of reasons why the pterosaurian characterisation of Piksi is problematic however: in short, its morphology is all wrong for a flying reptile and we suspect a non-pterosaurian ID is more likely. The Hornby specimen is thus a contender for the smallest latest Cretaceous pterosaur currently known.

A 1.5 m wingspan azhdarchoid next to one (SI) MrTiddlesmetre. From Martin-Silverstone et al. (2016).
This million dollar question, of course, is whether the specimen is a small juvenile or a small adult. The former would be neat, but the latter is potentially significant. The findings of recent, detailed histological examinations of pterosaur fossils are permitting increasingly good understanding of their growth regimes (e.g. de Ricqles et al. 2000; Prondvai et al. 2012), so we made a section of the humerus to understand how old the Hornby animal was when it died. Our section showed a mix of bone textures, some indicating that the specimen was still growing, but other features (secondary osteons, an endosteal lamella, lines of arrested growth and a large structure forming on the internal bone surface) are indicative of relative maturity (de Ricqles et al. 2000; Prondvai et al. 2012). We found the endosteal lamella (a band of bone deposited around the internal bone cavity) of particular interest, as this seems to signify the end of internal bone expansion in azhdarchoids, and is thus a hallmark of near-mature animals (note that this is not true for all pterosaurs - see Prondvai et al. 2012). The fused dorsal vertebrae are a further marker of maturity, as pterosaurs do not develop these features until they're at least subadults. The exact timing of notarium formation seems to differ from taxon to taxon (e.g. Bennett 1993; Kellner 2015), but their development does not seem to start until these animals were near to full size, if not at full size already. Putting these and a few other observations together suggests that the Hornby pterosaur was a latest-stage juvenile or subadult: in other words, it looks like a genuinely small pterosaur, not just a juvenile one. We don't know how much larger it might have got before it reached full size, but its ontogenetic characteristics and what we know of pterosaur growth regimes suggests it was close to maximum size at time of death. Given its estimated 1.5 m wingspan, it had a good chance of remaining smaller than the next smallest, 2.5 m wingspan pterosaur currently known the Campanian or Maastrichtian (McGowen et al. 2002).

What's inside the RBCM.EH.2009.019.0001 humerus? A mix of things, but among them are features indicative of late-stage juvenility/subadulthood. Please see the paper for details of this figure. From Martin-Silverstone et al. 2016.

A small pterosaur amongst the pigeons

There's obviously a limit to what a single fragmentary specimen can tell you about the evolution of a group, but what the Hornby specimen means for pterosaur evolution is interesting and - if we've interpreted it correctly - potentially significant. Most obviously, it suggests that small pterosaurs may have been present in the Campanian stage of the Late Cretaceous after all, at least in one part of the world. Regular readers will be aware that there's growing evidence for Late Cretaceous pterosaur faunas being less uniform than previously realised (e.g. Vremir et al. 2013, 2015), and our new specimen plugs into this picture nicely: it increasingly seems that the end Cretaceous wasn't just a stage for large-to-giant long-necked azhdarchids. What's more, while the specimen only provides one data point against the idea that birds ousted small pterosaurs, the presence of at least two types of bird in the Northumberland Formation seems to indicate small pterosaurs and birds coexisted in at least this palaeoenvironment. We might see this as a continuation of the coexistence pterosaurs and birds demonstrate in Jurassic and Early Cretaceous localities: maybe pterosaurs and birds got along OK after all.

...except when pterosaurs stole their eggs. Our PR art for the new paper, where a group of Hornby azhdarchoids perform guerrilla raids on shore-living Campanian bird nests. Take THAT, birds.
To my mind, one of the most significant things we do in the paper is discuss the 'face value' interpretations of Late Cretaceous pterosaur diversity: should we really be interpreting the lack of small pterosaur fossils as a genuine feature of their history when their fossil record is so patchy? We point out that some types of small pterosaurs - juveniles - had to exist in the Late Cretaceous, and yet their fossils are almost entirely unknown. We argue that this indicates a preservation bias against small bodied pterosaurs of any kind in the Campanian and Maastrichtian. Until we amass a good number of small juvenile pterosaur bones from this time without any small adults we cannot distinguish preservational interference from genuine biological signals. Perhaps the shift of pterosaurs from marine to non-marine habits through the Cretaceous (Butler et al. 2013) accounts for this lack of data. It's well known that terrestrial settings are less conducive to preserving relatively delicate fossils and small examples of even robust terrestrial animals like dinosaurs rarely fossilise in these deposits. We have to wonder what chance small pterosaur skeletons - which were strong in life, but fragile and weak once exposed to decay - have of making it into the fossil record in these settings. The fact the Hornby specimen is in such a sorry state perhaps reflects the rough time small pterosaur fossils experience under 'typical' fossilisation regimes, rather than the far gentler handling of animal remains evident at fossil Lagerstätten.

With all this said, the most important message of the paper has to be this: we need more data on small pterosaurs in the latest Cretaceous. The specimens we have are scrappy, hard to work with and offer limited scope for analysis. Thus, any small Late Cretaceous pterosaur material is significant, and whether they're lying unnoticed in museum collections or pulled straight out of the field, they are noteworthy specimens which need to be put on record. Curators and researchers, please keep your eyes peeled!

And that, in a nutshell, is our new paper: be sure to check it out if you want more details. You can also read Liz's take on the study over at The Conversation and other experts have been chiming in at news sites covering the story. With a bit of luck, this is not the only news you'll be hearing about Late Cretaceous pterosaurs from these quarters this year - more on these projects as they move along. All that's left to do is to thank Liz and Victoria for inviting me to collaborate with them on the new specimen - I learned a huge amount trying to get my head around this challenging material and its histology, and had a blast working with them.

This blogpost, paper and artwork are sponsored by Patreon

Regular readers will know that this blog and its art are sponsored by a suite of awesome Patrons, but this post is proof that this support goes further than mere internet tomfoolery and contributes to papers and outreach, too. Supporting my blog from $1 a month not only helps keep this blog ticking over, but helps me contribute thoughts, words and illustrations to scientific research. In return you get access to bonus blog content: additional commentary, in-progress sneak-previews of paintings, high-resolution artwork, and even free prints. For this post, we'll be talking about the PR art I've done for this research: how was the azhdarchoid reconstructed from that pile of rubble specimen? What's the story with the landscape image and why is there an ammonite in it? How many iterations did we go through to get that composition? Head over to Patreon to get access to this and the rest of my exclusive content!


  • Agnolin, F. L., & Varricchio, D. (2012). Systematic reinterpretation of Piksi barbarulna Varricchio, 2002 from the Two Medicine Formation (Upper Cretaceous) of Western USA (Montana) as a pterosaur rather than a bird. Geodiversitas, 34(4), 883-894.
  • Bennett, S. C. (1993). The ontogeny of Pteranodon and other pterosaurs. Paleobiology, 19(01), 92-106.
  • Benson, R. B., Frigot, R. A., Goswami, A., Andres, B., & Butler, R. J. (2014). Competition and constraint drove Cope's rule in the evolution of giant flying reptiles. Nature communications, 5, 3567.
  • Butler, R. J., Benson, R. B., & Barrett, P. M. (2013). Pterosaur diversity: untangling the influence of sampling biases, Lagerstätten, and genuine biodiversity signals. Palaeogeography, Palaeoclimatology, Palaeoecology, 372, 78-87.
  • Godfrey, S. J., & Currie, P. J. (2005). Pterosaurs. Dinosaur Provincial Park: A Spectacular Ancient Ecosystem Revealed, 292-311.
  • Hone, D. W. E., & Benton, M. J. (2007). Cope's Rule in the Pterosauria, and differing perceptions of Cope's Rule at different taxonomic levels. Journal of Evolutionary Biology, 20(3), 1164-1170.
  • Kellner, A. W. (2015). Comments on Triassic pterosaurs with discussion about ontogeny and description of new taxa. Anais da Academia Brasileira de Ciências, 87(2), 669-689.
  • Martin-Silverstone, E., Witton, M. P., Arbour, V. M, & Currie, P. J. (2016). A small azhdarchoid pterosaur from the latest Cretaceous, the age of flying giants. Royal Society Open Access, 3, 160333.
  • McGowen, M. R., Padian, K., De Sosa, M. A., & Harmon, R. J. (2002). Description of Montanazhdarcho minor, an azhdarchid pterosaur from the Two Medicine Formation (Campanian) of Montana. PaleoBios, 22(1), 1-9.
  • Prondvai, E., Stein, K., Ősi, A., & Sander, M. P. (2012). Life history of Rhamphorhynchus inferred from bone histology and the diversity of pterosaurian growth strategies. PLoS One, 7(2), e31392.
  • Vremir, M., Kellner, A. W., Naish, D., & Dyke, G. J. (2013). A new azhdarchid pterosaur from the Late Cretaceous of the Transylvanian Basin, Romania: implications for azhdarchid diversity and distribution. PLoS One, 8(1), e54268.
  • Vremir, M., Witton, M., Naish, D., Dyke, G., Brusatte, S. L., Norell, M., & Totoianu, R. (2015). A Medium-Sized Robust-Necked Azhdarchid Pterosaur (Pterodactyloidea: Azhdarchidae) from the Maastrichtian of Pui (Haţ eg Basin, Transylvania, Romania). American Museum Novitates, (3827), 1-16.

Friday, 12 August 2016

Trunk or no trunk, small or giant ears, long or short neck... what did the giant rhinocerotoid Paraceratherium really look like?

Giant, Oligocene rhinocerotoids Paraceratherium transouralicum engage in some early morning flirting. Because, in rhino speak, playing hard to get involves shoulder barges and head-butts.
Depictions of the giant indricotherines, relatives of modern rhinoceros that lived across mid- and eastern Asia during the Oligocene, have varied over time. We've known about these animals - which are part of a longer-lived (Eocene-Miocene) indricotherine lineage that includes a number of smaller, almost okapi-or horse-like species - for over 100 years and they have become regular fixtures in museums, books and those rare documentaries which offer glimpses into ancient life outside of the Mesozoic. Any yet, when we think of our favourite indricothere paintings - including those by our most celebrated mammalian palaeoartists such as Knight, Burian, Anton, and Buell - they often differ markedly in their depiction of these 15-20 tonne animals. Most notably, their neck proportions, overall robustness, the development of a proboscis or trunk, and - most recently - the size of the ears are all inconsistent. Why are these animals so differently depicted, and should we rule out some of the anatomies we've seen in palaeoart in the last century? Having faced these questions recently when asked to restore this animal myself (above), I thought I'd share some of what I learned in my research here.

The obligatory note on nomenclature

It almost seems tradition that any article or paper on indricotherines requires an aside on their confused taxonomy. As has been the case for decades now, the taxonomy and systematic nomenclature of these giant rhinocerotoids are a matter of ongoing discussion. It is widely appreciated that several giant indricotherine species from roughly contemporaneous Oligocene Asian sediments can be identified, but how many species they represent, and how they are related to each other, is not clear. At least seven generic titles and many more species names have been given to the largest of these animals over the years (Indricotherium and Baluchitherium are perhaps the most famous generic labels), but some authors (e.g. Lucas and Sobus 1989, Prothero 2013) tidy all or most of these taxa into three species of the oldest established genus, Paraceratherium. Arguments persist, however, that at least one other, perhaps slightly smaller genus existed, Dzungariotherium (Qiu and Wang 2007). Geologically older indricotherine genera such as the Eocene Urtinotherium are also wrapped into these discussions as remains attributed to the Oligocene genera are sometimes argued as having greater affinity to these older taxa (Prothero 2013).

This confusion is sometimes framed as a 'lumper/splitter' philosophical distinction, but it does not help that the fossil record of these giant rhinocerotoids is far from exemplar: giant indricotherine specimens can be fragmentary, of starkly contrasting size with one another, and many suffer from distortion. The fact that 20th century indricotherine science developed with Asian and American teams working largely in isolation, with limited access to certain specimens and literature, has also contributed to the confused history of this group. Those interested in the history of indricothere taxonomy should check out Prothero (2013) for an overview. For now, it will serve us to simply state that the best known, biggest and most famous of these animals currently resides as the taxonomic address of Paraceratherium transouralicum. This is the species most of us think of as 'the' giant indriotherine as well as the taxon that has carried both the Indricotherium and Baluchitherium label at one time or another. It's also the focus of most artwork of giant rhinoceratoids, and thus forms our primary interest here.

Giant rhino, bulky giraffe or giant workhorse?

One reason we see such variation in indricotherine appearance is that researchers have produced vastly different interpretations of its anatomy in the last 100 years. But unlike, say, dinosaurs, where older reconstructions have been (for the most part) abandoned in favour of newer, more accurate interpretations, educators and researchers continue to publish skeletal reconstructions published in the 20s and 30s despite our improved knowledge of indricotherine anatomy, documented criticisms of these older works, and the availability of more modern, theoretically better-informed reconstructions.

Many readers may be aware that the first reconstruction of Paraceratherium, published by Osborn (1923a), showed a form not too far off a giant rhinoceros - a heavyset, short-necked animal with a deep torso and short legs. Osborn published a revised version almost immediately after his first effort, which had a much longer neck and longer legs thanks to data provided by additional fossil material (Osborn 1923b). A shorter-necked version was then produced by Granger and Gregory (1935, 1936), who scaled the remains of numerous, differently-sized individuals from a range of collections to create their robust, gigantic take on indricotherine anatomy. Although this reconstruction has been quite influential, Fortelius and Kappelman (1993) have been critical of the scaling methods used by Granger and Gregory, calling their interpretation 'a highly speculative creation indeed'.

Paraceratherium has been variably reconstructed over the years, with particular disagreement over how long the neck was compared to the body. So far as I can tell, a consensus on the life appearance of these animals has yet to be reached.
A third contrasting reconstruction was published a few decades later by Gromova (1959), based on a composite mounted skeleton in the Paleontological Institute, Russian Academy of Sciences. This reconstruction, executed by N. Yanshinova, was accompanied by several wonderful muscle and skin reconstructions which palaeoart fans will not want to miss. Both the mount and reconstruction show a gracile, giraffe-like form with a remarkably long neck and, in being based on a relatively complete set of giant indricotherine remains, some have argued it is a superior take on indricotherine anatomy than those produced by Osborn, or Granger and Gregory (Fortelius and Kappelman 1993). The most striking aspect of this reconstruction is its very long neck. We have to stress that this is extrapolated from a few incomplete cervicals associated with postcranial material, and its exact length remains uncertain - a complete set of neck bones remains elusive for Paraceratherium. This is another reconstruction which has been quite influential (helped, no doubt, by its apparent basis for the BBC's Walking with Beasts 'Indricotherium') but, again, it has not escaped criticism. Paul (1997) suggested that multiple aspects of this mount and reconstruction were erroneous, including the length of the neck, the size of the pelvis and depth of the ribcage, the length of the feet, and the ratio of the humerus and femur, as well as the fully erect posture of the limbs.

And so we turn to another indricotherine skeletal reconstruction, produced by Paul (1997). This restoration incorporated data from the same specimens used in the efforts above and came out somewhat 'averaged' between the more heavyset restorations of the early 20th century and the gracile interpretation of the 1950s. It looks, in overall form, more like a giant workhorse than it does a giant rhino or bulky giraffe. Paul (1997) provides some discussion of the reconstruction process - this is worth a read if you're interested in the life appearance of Paraceratherium and its relatives. Paul's interpretation has, to my knowledge, escaped criticism to date and, to the contrary, Larramendi (2016) described this reconstruction as 'accurate', although did not elaborate on why it should be considered superior to older efforts.

The million dollar question here is obvious: which one of these different takes on Paraceratherium is 'right'? To be honest, I'm not sure. The situation is compounded by the fact that a lot of indricotherine literature is obscure, that the specimens fragmentary and that many of them await description. I was hoping that Donald Prothero's recent (2013) book Rhinoceros Giants, which is solely dedicated to Paraceratherium, would provide some insight on this matter, but it's not a great help here - it provides no real evaluation on the different reconstructions and does not even mention Paul's 1997 effort. My work above is primarily based on Paul's (1997) skeletal but this is largely because of principle rather than real insight. Paul's work is the most modern and, of course, he's made a career out of reliably reconstructing extinct animals. The brief endorsement from Larramendi (2016) helps here too, of course, but a longer discussion of the relative merits and detriments of each interpretation would be useful. Opinions from others with more insight into this matter are welcome in the comments below.

A tapir-like proboscis... on a rhino?

Turning our attention to the face, did Paraceratherium and its relatives have relatively short-lipped faces like those of rhinos, or long, mobile proboscides like their more distant relatives, the tapirs? Despite mammal lips and nasal tissues being highly fleshly and thus only rarely entering the fossil record, this is a surprisingly easy question to answer. Whether rhino, tapir or anything else, a suite of osteological characters seem to correlate well with the presence of proboscides. Briefly summarised, these are: narrow snouts; retraction of the nasal openings towards the orbits; the presence of large muscle scars, bony knobs and other muscle attachment markers around the nasal opening (particularly in the dorsal region); retraction of the nasal bone (the 'roof' of of the nasal opening); deepening of the premaxillary bone (the bone making the jaw tip); anterior migration of the orbit; a large intraorbital canal (a foramen situated in the cheek region, just in front of the eye - it houses the nerves and blood vessels for our anterior face muscles); and strengthening of the posterior skull regions related to supporting the weight of the head on the neck (Wall 1980). Note that the criteria for elephant-like trunks are similar, but slightly different.

Paraceratherium transouralicum (formerly Baluchitherium grangeri) skull in dorsal, lateral and ventral views. Note features around the skull anterior linked to proboscis development (see text). From Osborn (1923b).
Paraceratherium skulls (above) meet these criteria well and, all else being equal, we have to say that yes, it looks likely that these giant rhinoceratoids had short proboscides in life, presumably to assist browsing from trees and bushes (Prothero 2013). The view that they had more typically rhinoceros-like faces is hard to defend in light of these cranial features: mammal skulls just don't have those retracted nasal openings, associated deep muscle scarring etc. unless they were doing something unusual and sophisticated with their upper lip and nasal tissues. The reality of giant indricotherines with dangly noses may seem hard to swallow for those of us used to shorter lipped versions, but given the relationships between rhinos and tapirs, the fact that some other fossil rhinocerotoids probably had proboscides as well (e.g. Wall 1980), and the independent development of long, flexible noses in numerous mammal lineages, we can't really see this as unusual. Moreover, we need to remember that modern rhinos are derived animals in their own right and separated from the indricotherine lineage by tens of millions of years. They aren't necessarily always going to be the best models for the life appearance of their fossil ancestors.

And big, elephant-like ears, right?

Finally, let's tackle the component that everyone now mentions about indricotheres since seeing the Carl Buell's cover art for Donald Prothero's Rhinoceros Giants:

Indiana University Press.
Yikes, elephant ears? For those of us familiar with the history of indricotheres in art, where their ears are restored as typically rhinoceros-like, this is a shocking, double-take image. Within the book, Prothero justifies the restoration:

"...indricotheres were larger in body mass than any living elephant and almost certainly had problems regulating their body heat at such large size. Elephants must do all they can to increase the surface area of their bodies to release as much excess heat as possible, which is why they have huge fan-like ears full of blood vessels that are essentially giant radiators. Given the huge size of indricotheres, it seems likely that they too should have had elephant-like ears, or at least very large ears of some shape, much larger than they are usually drawn."
Prothero, 2013, p. 90.

The text continues to suggest that this appearance is not without anatomical support, the prominence of the mastoid and paroccipital processes (projections of bone situated behind the ear opening, adjacent to the posterior surface of the skull) being similar to the condition in certain elephants and mastodonts, and therefore indicative of large, flappy ears (Prothero 2013).

I have mixed feelings about this reconstruction. I like it for two reasons. The first is that it's nice to see indricotheres being distanced from their depiction as giant, long-necked rhinoceroses - again, it's not unreasonable to think they may have looked quite different in to modern rhinocerotids in many aspects. I also like these ears for being an All Yesterdays-style speculation on soft-tissue adaptations in extinct species. If we can use this as an excuse to give fat stores to desert sauropods or fuzzy hides to Arctic ceratopsids, then we can give large ears to giant rhinoceratoids.

On the other hand, I'm not convinced that they're as likely as Rhinoceros Giants suggests. It's clear from our modern fauna that ear size does not correlate with body mass in terrestrial mammals. By this logic many rhinos and giraffes should have proportionally large ears too, which they evidently do not. We also have to consider that even larger animals than indricotheres, dinosaurs, almost certainly got by without giant ears to help lose heat. And yes, while dinosaurs may have used different metabolic strategies to mammals, one inescapable consequence of giant size is a constant high body temperature. At least some investigations into the proportions of large dinosaurs suggest that development of their features - such as sauropod necks - were not driven by thermoregulatory pressures (Henderson 2013).

We should also consider the unusual nature of elephant thermoregulation: they are not typical mammals when it comes to controlling body heat. For one, they're atypically compact compared to other large mammals because they have extremely short necks, giant, round heads, and big, rotund torsos. This is a suboptimal bauplan for thermoregulation because it minimises surface area with respect to volume, and thus reduces the available area for elephants to dump excess heat. Moreover, unlike most mammals, they lack sweat glands (Wright and Luck 1984), do not pant, and they live in climates which are so warm that for much of the day they cannot shed heat through simple convection, big ears or not (Weissenböck et al. 2012). Elephants can, of course, regulate their temperature, but they need to employ different strategies to the rest of us mammals. These include maintaining moist skin with mud bathing and trunk spraying (Wright and Luck 1984), maintaining a sparse set of body hair to aid thermal escape (Myhrvold et al. 2012), using heterothermy (Weissenböck et al. 2012), the development of 'thermal windows' in their skin (Weissenböck et al. 2010), having loose and highly wrinkled skin to boost surface area and - of course - fanning their blood-vessel rich ears to help lose heat, when ambient temperatures are low enough for this to make a difference.

Silhouettes of the largest land mammals of all time, Paraceratherium transouralicum and Palaeoloxodon namadicus. Note the relatively gracile build of Paraceratherium - all the better for improving surface area:volume ratio, and thus superior for radiating heat. The numbers at the base of the image refer to estimated shoulder heights and tonnage. From Larramendi (2016).
These facts suggest elephants should not be used as direct thermoregulatory models for a giant rhinoceratoid. Modern rhinos other perissodactyls are much more typical in their thermoregulatory approaches: they have sweat glands and use panting behaviours (Hiley 1977) as well as some special tactics, such as enhanced vascularisation in the skin folds of certain rhino species (Endo et al. 2009). We have to assume that indricotherines at least had these entry level perissodactyl adaptations and, if so, they would have an advantage over elephants in hot climates. Indricotherines also benefit from being more complicated in form than elephants. They have longer limbs and necks, as well as a proportionally smaller head, and this enhances their surface area:volume ratio. Again, makes them better adapted to cope with heat as they have a shape better suited to radiating excess body heat. And of course, there's no reason to assume this could not have been augmented with wrinkled or folded skin or sparse hair. The picture emerging from these points is that big ears are only one strategy that big animals may use to keep cool, and maybe one that will only arise in specific circumstances. The idea that indricotherines would have big ears just because of their size is far from certain.

Basic muscle layout and trajectories (arrowed lines) of a modern horse. Note their superficial attachment and position high on the head - the ear canal itself is about halfway down the back of the skull. The 's' is the scutiform cartilage, which hangs out in front of the ear over the jaw muscles. From Goldfinger (2004).
But isn't all this moot because of Prothero's (2013) observations about the mastoid and paroccipital processeses being expanded, and thus giving big ears something to hang off? I'm suspicious about the significance of this observation. So far as I can determine, neither the mastoid or paroccipital have anything to do with anchoring ear tissues in modern perissodactyls (or perhaps any mammal). This might be because in most mammals - primates being one obvious exception - the ear pinnae are vertically displaced from the ear canal and attach to the head via a series of muscles and cartilages at the top of the skull (above). Only select few of the ear muscles reach the skull directly and these anchor, with very small attachments, to the skull midline, dorsoposterior margin and zygomatic arch. The rest have no osteological connection at all, anchoring instead to cartilage, membranes overlying facial musculature, or even the side of saliva glands. The paraoccipital and mastoid processes do have important roles in the muscular system but these are to do with neck, jaw and tongue muscles, not ears. Thus, unless indrictotheres were doing something different to modern mammals, those particularly big processes behind their ear openings were probably more to do with supporting and moving the head than they were holding big ears, and may have little significance to the big-eared indricotherine hypothesis.


Putting all this together, it seems that there might be less need for uncertainty about indricothere appearance than our various artworks suggest. We should be saying 'yes' to some sort of proboscis, and 'probably not' to big ears (or, at least, 'there's no reason for them'). The elephant (or, giant rhino, if you prefer) in the room is the proportion issue, and it would be good to see folks who really know rhinocerotoid anatomy pore over those various reconstructions to ascertain which (if any) are the best representation of indricotherine form.

Next time: either the Next Big (but also kinda small) Thing in pterosaur research, or another trip to the Triassic.

Big rhinos need big support - thank goodness for Patreon

The paintings and words featured here are sponsored by another group of (metaphorically) giant mammals, my Patreon backers. Supporting my blog from $1 a month helps me produce researched and detailed articles with paintings to accompany them, and in return you get access to bonus blog content: additional commentary, in-progress sneak-previews of paintings, high-resolution artwork, and even free prints. For this post, we'll be taking a further look at the anatomy of the Paracertherium in my painting, above. Why do they have little manes and stripy faces? Are those child rhinos at the back a bit fuzzy? And why do the main animals look like they're fighting? Head over, and sign up to Patreon to get access to this and the rest of my exclusive content!


  • Endo, H., Kobayashi, H., Koyabu, D., Hayashida, A., Jogahara, T., Taru, H., Oishi, M., Itou, T., Koie, H. & Sakai, T. (2009). The morphological basis of the armor-like folded skin of the greater Indian rhinoceros as a thermoregulator. Mammal Study, 34(4), 195-200.
  • Fortelius, M., Kappelman, J., 1993. The largest land mammal ever imagined. Zoological Journal of the Linnean Society, 108, 85-101.
  • Goldfinger, E. (2004). Animal Anatomy for Artists: The Elements of Form. Oxford University Press.
  • Granger, W., & Gregory, W. K. (1935). A revised restoration of the skeleton of Baluchitherium, gigantic fossil rhinoceros of Central Asia. American Museum of Natural History, 787, 1-3.
  • Granger, W., & Gregory, W. K. (1936). Further notes on the gigantic extinct rhinoceros, Baluchitherium, from the Oligocene of Mongolia. American Museum of Natural History, 72, 1-73.
  • Gromova, V. (1959). Giant rhinoceroses. Trudy Paleontologiskei Institut Akademie Nauk, 71, 1-164.
  • Henderson, D. M. (2013). Sauropod necks: are they really for heat loss?. PloS one, 8(10), e77108.
  • Hiley, P. G. (1977). The thermoregulatory response of the rhinoceros (Diceros bicornis and Ceratotherium simum) and the zebra (Equus burchelli) to diurnal temperature change. African Journal of Ecology, 15, 337-337.
  • Larramendi, A. (2016). Shoulder height, body mass and shape of proboscideans. Acta Palaeontologica Polonica, 61, 537-574
  • Lucas, S. G., & Sobus, J. C. (1989). The systematics of indricotheres. In: Prothero, D. R., and R. M. Schoch (eds.) The Evolution of Perissodactyls. Oxford University Press, New York, 358-378.
  • Myhrvold, C. L., Stone, H. A., & Bou-Zeid, E. (2012). What is the use of elephant hair?. PloS one, 7(10), e47018.
  • Osborn, H. F. (1923). The extinct giant rhinoceros Baluchitherium of Western and Central Asia. Natural History, 23, 208–228.
  • Osborn, H. F., & Berkey, C. P. (1923b). Baluchitherium grangeri, a giant hornless rhinoceros from Mongolia. American Museum of Natural History, 78, 1-15.
  • Qiu, Z. X., Wang, B. Y., 2007. Paracerathere fossils of China. Palaeontologia Sinica, C29, 1-396
  • Paul, G. S. (1997). Dinosaur models: the good, the bad, and using them to estimate the mass of dinosaurs. DinoFest International Proceedings. Philadelphia: The Academy of Natural Sciences, 129-154.
  • Prothero, D. R. (2013). Rhinoceros Giants: The Paleobiology of Indricotheres. Indiana University Press.
  • Wall, W. P. (1980). Cranial evidence for a proboscis in Cadurcodon and a review of snout structure in the family Amynodontidae (Perissodactyla, Rhinocerotoidea). Journal of Paleontology, 54, 968-977.
  • Weissenböck, N. M., Weiss, C. M., Schwammer, H. M., & Kratochvil, H. (2010). Thermal windows on the body surface of African elephants (Loxodonta africana) studied by infrared thermography. Journal of Thermal Biology, 35, 182-188.
  • Weissenböck, N. M., Arnold, W., & Ruf, T. (2012). Taking the heat: thermoregulation in Asian elephants under different climatic conditions. Journal of Comparative Physiology B, 182(2), 311-319.
  • Wright, P. G., & Luck, C. P. (1984). Do elephants need to sweat?. South African Journal of Zoology, 19(4), 270-274.