Palaeoartworks, the case studies, part 2: Feathered dinosaurs and tiny Crocodyliformes

It's time for part 2 of our 'Palaeoart Case Studies' series, this time featuring two subjects: the tiny Cretaceous crocodyliform Koumpiodontosuchus and the probable Lower Cretaceous troodontid Yaverlandia bitholus. Unlike our last subject in this series, giant pterosaurs, neither of these animals is huge. Koumpiodontosuchus is particularly diminutive with an estimated adult length of 600 mm long. Presenting the scale of an animal accurate is important for good palaeoart, as it's not only an important factor in the animal's biology and ecology, but also an integral part of it's character. Small animals present palaeoartists with a particular challenge because many folks hold the preconception that all extinct animals were large. For diagrammatic images, simply adding a person, modern animal or familiar object next to our creature shows its size, but this cannot work when rendering scenes that took place millions before familiar entities appeared. How do palaeoartists get around this? Read on to find out.

Yaverlandia represents the result of a recent palaeoart success story. After many years of trying, it seems palaeoartists have finally got the hang of recreating convincing-looking feathered dinosaurs. How did they do this? In short, by abandoning the need to show all dinosaurs as scaly reptiles and embracing the birdiness inherent to many species (or, to flip this around, realising that many traits unique to modern birds were common to many of their dinosaur ancestors). But what does this mean for the way we reconstruct troodontids and other feathered dinosaurs? Again, the answers are below.

This series of case studies is in aid of my art gallery in Lyme Regis, running until May 4th, at the Town Mill. Full details here.

Koumpiodontosuchus: a tiny, button-toothed crocodyliform

Reconstruction of the tiny Wealden bernissartid Koumpiodontosuchus aprosdokitii. Hopefully, you can see that it's not a large animal without having to think about it too much, but why is that? Prints of this image are available here.

The size of extinct animals a favourite topic of palaeontology aficionados, and presenting it accurately is an important goal for palaoartists. Although we often think of extinct species as very large animals, most were not giants. The skull of the tiny Cretaceous crocodyliform Koumpiodontosuchus show below, for instance, belonged to an adult individual that, in life, was about 600 mm long. But how can palaeoartists express a sense of animal size - big or small - without the use of modern objects, animals or people for reference?

Some general trends of animal appearance can be useful in conveying size in extinct species. These probably aren’t features that most of us think about when observing animals, but they provide palaoartistists with some tricks to give a sense of scale to their subject matter without using other objects or animals for scale. Generally speaking, facial features - particularly the eyes - of larger creatures are relatively smaller than those of more diminutive animals. The limb bones of larger animals are more robust and, as they approach the extremities, are proportionally shorter. Smaller animals often have less conspicuous muscle contours than larger animals, particularly if they have a fluffy covering and, in being lighter, smaller creatures are frequently more sprightly and ‘weightless’ than larger ones.

Animal proportions only give a very general sense of scale, however. To give a more precise measure, palaeoartists often juxtapose relatively familiar species alongside their subjects. The use of ‘background’ animals, or different varieties of plants, are useful in this respect. Even if the audience is not very familiar with these background entities, their proportions in relation to the subject gives an impression of scale. This can work in inverse, too: bigger animals or plants, shown in low contrast at a distance, can help reinforce the size of smaller subjects. Crafty consideration of point of view can also help: does the subject need a very low, tight point of view to be seen, or does a wider frame capture it more adequately? Again, these are not necessarily factors that we consider when viewing animals or artwork of them, but they are essential considerations for palaeoartists attempting to reconstruct not only the anatomy and lifestyle of their subject, but also their size and physical presence.

Yaverlandia: Britain's most bird-like dinosaur

Two Yaverlandia investigate a termite-riddent tree stump in Lower Cretaceous Britain. As with many other dinosaurs, their depiction here as very bird-like creatures is the result of palaeoartists having to completely overhaul the way feathered dinosaurs are rendered. 

Known only from skull caps, Yaverlandia bitholus is one of Britain’s lesser known dinosaurs. Found in Lower Cretaceous Wessex Formation cliffs close to its namesake, the Isle of Wight coastal village of Yaverland, it was thought for a long time to represent a member of the bone-headed dinosaur group Pachycephalosauridae. Recently, it has been reinterpreted as Britain’s first troodontid, an omnivorous theropod dinosaur closely related to birds and familiar carnivores like Troodon, Velociraptor and Deinonychus. Troodontids were big brained, nimble animals and most were rather small. Yaverlandia was no exception, with a tentative length estimate of about 2 m. 

Any restoration of Yaverlandia and its kin has to incorporate data on dinosaur feathering, details of which have only been available in earnest for the last 20 years or so. The uptake of feathered dinosaurs among palaeoartists has been variable. Some restrict them to specific regions of the body, while others provide smatterings of feathers across primarily scaly skin. Increasing numbers of artists restore these animals as very bird-like however, with feathers across their entire bodies including their heads, tails and legs. Fossil evidence is clearly in line with this latter approach. Bird-like dinosaurs, including troodontids, have been discovered with extensive feathering that covers their not only their torsos necks, heads and tails, but even sometimes their legs and toes. 

These discoveries have given palaeoartists a lot to think about when restoring the appearance of bird-like dinosaur species. Feathers are complex, three-dimensional structures which alter the body profile of their owners. They plump the appearance of the body and smooth contours of the animal’s profile. Thus, if a dinosaur has complex feathers, we can no longer simply restore their musculoskeletal system and wrap skin over it, as often done in the past. What’s more, palaeoartists have to think carefully about the way feathered dinosaur arms and hands are rendered, because their feather configuration is similarly complex modern bird wings: feathers erupt from the tip of the second finger to the elbow, with feathers of the shoulders covering the ‘gap’ between forelimb and body feathering. 

These ‘new look’ dinosaurs are more bird-like than ever, and evidence is mounting that feathers appeared much earlier in dinosaur evolution than classically realised. It is even quite probable that Tyrannosaurus was feathered. This is disheartening to some who ‘prefer’ the appearance of scaly dinosaurs, and many still erroneously render troodontids and their kin with scaly hides. Wholly feathered restorations of such dinosaurs are without doubt factually accurate renditions however, and entirely uncontroversial among scientists. 

Episode 3: Bernissartids, the button-toothed Crocodyliformes

3/3 - this, ladies and gentlemen, is the end. At least, until the inevitable prequels where I'll ignore the canon of the expanded universe and do my best to tarnish everything you liked about the original trilogy.

Here we are then, the last instalment of the Wealden Crocodyliformes Trilogy. Following the posts on atoposaurids and goniopholidids, today we're going out with a bang by covering a newly described Wealden crocodyliform unleashed on the world this morning. The study was written up by my University of Portsmouth chums and colleagues Steve Sweetman, Ulysse Pedreira-Segade and Steven Vidovic (Sweetman et al. 2014), and Steve V. has covered some aspects of his involvement at his blog. The paper is open-access so, for the full skinny on the discovery, you should head here.

This most recently identified Wealden crocodyliform is among the most sophisticated and unusual of all Wealden crocs. Named Koumpiodontosuchus aprosdokitii, it is known from a well-preserved skull which was recovered in circumstances owing much to chance and good fortune (Sweetman et al. 2014). This animal is currently only known for certain from the Wessex Formation of the Isle of Wight, specifically from fossil-rich cliffs next to the seaside village of Yaverland, and the only known skull of it is broken in half. The posterior half was discovered in March 2011 by holidaying fossil hunters, who took it to the local dinosaur museum (Dinosaur Isle, of Sandown) to have it identified. Another family, on a fossil-hunting holiday three months later, then found the front half of the skull. They took this to the same museum where, by chance, the same museum staff who’d handled the first piece were on hand. It was realised that each piece belonged to the same specimen, and the first half was rapidly brought back to the museum to check the degree of articulation. Remarkably, the join between the broken pieces was near perfect – clearly neither chunk had been exposed to weathering effects very long before being discovered – and the entire skull could be seen. Each piece was then donated to the museum to allow its study. Given the chain of events and people involved in the discovery of Koumpiodontosuchus, it’s easy to imagine how only single halves of the skull might be known to science, or even neither. This is clearly yet another story which stresses the importance of amateur fossil hunters to Wealden fossil discoveries, and the benefits of responsible collecting.

Holotype skull and mandible of the button-toothed crocodyliform, Koumpiodontosuchus aprosdokitii. From Sweetman et al. 2014.

Button-toothed crocodiles in context

Koumpiodontosuchus is a member of Bernissartidae, a group named by Sweetman et al. (2014) which only contains two species: Koumpiodontosuchus and Bernissartia fagesii. The latter is a famous, small Jurassic and Cretaceous crocodyliform known from France, Denmark, Spain, Portugal and particularly Belgium, where a spectacular complete skeleton has been unearthed. Indeterminate species of Bernissartia also seem to occur in the Ashdown Formation of Hastings (Salisbury and Naish 2011), but this identification may eventually warrant reappraisal now that Koumpiodontosuchus has been discovered. Bernissartid remains are not new, some of the first material of these animals being documented in the 1850s and Bernissartia itself being named from Belgian fossils in the 1880s. Isolated teeth, likely referable to Koumpiodontosuchus, have been found in Wealden deposits since at least the 1970s (Buffetaut and Ford 1979), so were clearly present across the entire geographic and stratigraphic range of the Wealden Supergroup.

Bernissartia has long been a bit of an oddball among Crocodyliformes, possessing some unusual anatomy and being of uncertain placement in crocodyliform systematics. The discovery of Koumpiodontosuchus provided a bit of light on this front, suggesting that Bernissartia was part of a group containing at least one other similar species, and that they occupy an evolutionary place between atoposaurids and the goniopholidid + Eusuchia radiation. This position isn’t too surprising, as there are a number of features in bernissartids which link them to Eusuchia – see below. Bernissartidae is primarily defined by dental characteristics, with the most obvious one also being the namesake of Koumpiodontosuchus: “button-toothed crocodile” (if anyone wants a common name for these Crocodyliformes, this is the one to use). The posterior teeth of bernissartids are rather globose – wide, short and blunt – and distinctive compared to the dentitions of most other Crocodyliformes. It’s these teeth which, even in isolation, betrayed the presence of bernissartids in the Wealden well before the more substantial Koumpiodontosuchus fossil was discovered. Their other teeth are quite different to this, however. The mid-region dentition is rather conical in shape; ‘pseudocanines’ erupt about 25 % of the jaw length from the jaw tip, and conical teeth emerge procumbently from the jaw tips themselves. Koumpiodontosuchus has two large pseudocanines on its lower jaw, which erupt so close to each other that they share a single, enlarged tooth socket. Bernissartia, by contrast, only possesses one.

The new Wealden bernissartid Koumpiodontosuchus aprosdokitii foraging for molluscs. It's eating a mud snail, Viviparus cariniferus, while tiny (6 mm long) physid gastropods Prophysa crawl over pond scum in the lower left of the image. Dragonflies provide scale, while unnamed tetanurans (based on findings of Benson et al. 2009) prowl around the background. An earlier version of this reconstruction was featured in Sweetman et al. (2014). Prints of this image are available here.

Bernissartids packed this sophisticated dentition into relatively tiny jaws: these were not big crocodyliforms. Indeed, with body lengths of approximately 600 mm, bernissartids were probably the smallest crocodyliform species in the entire Wealden succession. Like goniopholidids, bernissartids bore osteoderm shields on their backs and bellies, but the dorsal series was rather more complex than those of other Wealden crocodyliforms. Rather than possessing two rows of interlocking osteoderms as we saw in goniopholidids and atoposaurids, bernissartids possess four rows of osteoderms along their backs. These comprise two sets of rectangular, double-keeled scutes along the midline, and laterally bordering square osteoderms with single keels (Salisbury and Frey 2001). None of these interlocked, and – based on what we’ve discussed for other Wealden Crocodyliformes – it’s worth considering what impact this had on bernissartid locomotion. Rather than supporting their trunks with scutes, it seems that bernissartids developed procoelus trunk vertebrae (that is, vertebrae with centra extending into the corpus of the vertebra behind) to support their bodies when walking (Salisbury and Frey 2001). This feature, along with their relatively complex osteoderms, is shared with eusuchians and are some of the reasons why these animals have classically been allied to these Crocodyliformes. Of further interest here is the biconvex nature of the first bernissartid tail vertebra – this has further implications for their locomotion, which we’ll get to below.

The bit on palaeoecology

Ecologically, it seems that bernissartids had a preference for hard shelled prey. Their blunt posterior dentition has been labelled as ‘tribodont’ – literally meaning ‘crushing teeth’ – and, like slamming a couple of anvils together, are ideally shaped to crunch hard shells. Some confirmation of this idea is seen in the wear facets often seen on tribodont bernissartid teeth. Classically, their prey was largely considered to comprise molluscs such as the freshwater snails and clams populating Wealden streams and lakes (Buffetaut and Ford 1979). Recently, a broader diet has been postulated for bernissartids however, the logic being that hard shells are hardly restricted to molluscs even in freshwater settings (Sweetman et al. 2014). Insects and crayfish probably formed as much of their diet as molluscs, all of which were likely procured or extracted from soft-substrates with the procumbent anterior teeth. We should not forget the savage-looking pseudocanines of these animals however: these would be of little use against hard prey items, but may have allowed for spearing relatively soft-animals. Perhaps bernissartids are best viewed as rather opportunistic feeders, primarily taking hard-shelled prey but not turning their noses to other types of food when the opportunity arose.

If gastropods like this Wealden mud snail, Viviparus cariniferus, had nightmares, they contained bernissartids. 

Where was most of this prey caught? There is evidence that bernissartids were equally at home in water and on land. Their biconvex first tail vertebra suggests their tails were capable of considerable movement for providing burst propulsion through water and, unlike most other Wealden Crocodyliformes, their lack of interlocking osteoderms facilitated lateral trunk motion (Salisbury and Frey 2001). While compromising overall speed, this may have permitted greater amounts of manoeuvrability – ideal for pursing nimble, if relatively slow, aquatic arthropods. We’ve already mentioned that the reinforced trunk vertebrae of bernissartids would provide ample reinforcement for terrestrial locomotion, and their small size is relevant here as well. Like the small-bodied atoposaurids, and unlike the big goniopholidids, bernissartids had relatively small amounts of weight to lug around on land and could likely sustain long periods of terrestrial locomotion without tiring. It’s possible, therefore, that they found much of their prey on land as well as in water, perhaps enjoying the beetles, cockroaches and other tough-shelled terrestrial insects known to occur in Wealden deposits.

It’s worth pointing out that bernissartids may not be the only Wealden Crocodyliformes adapted for hard-shelled prey. The poorly known, 1.5 m long Wealden eusuchian Hylaeochampsa vectiana also has large posterior teeth ideal for smashing shelled prey (Clark and Norell 1992), although the dentitions of other hylaeochampsids are complex and it’s possible Hylaeochampsa had a very varied diet. As discussed for other Wealden Crocodyliformes, it’s likely that the size difference between the bernissartids and Hylaeochampsa would prevent too much overlap in prey preference: the latter may have been capable of eating large molluscs or even small armoured vertebrates, which were probably unavailable to bernissartids. There's lots more we could say here, but I'd best not - maybe Hylaeochampsa will warrant dedicated discussion at a later date.

The end

And I guess that's where we'll leave the Wealden Crocodyliformes for now. As alluded to above, there are other crocodyliform species and groups we could discuss, but they're generally less well known than the taxa we've covered across these posts and it would be difficult to discuss them in comparative depth. I hope you've enjoyed this series of themed posts and, if artwork of ancient Wealden animals is your thing, come back soon for a big announcement about an event related to just that.

References

• Benson, R. B., Brusatte, S. L., Hutt, S., & Naish, D. (2009). A new large basal tetanuran (Dinosauria: Theropoda) from the Wessex Formation (Barremian) of the Isle of Wight, England. Journal of vertebrate Paleontology, 29(2), 612-615.
• Buffetaut, E., & Ford, R. L. E. (1979). The crocodilian Bernissartia in the Wealden of the Isle of Wight. Palaeontology, 22(4), 905-912.
• Clark, J. M., & Norell, M. (1992). The Early Cretaceous crocodylomorph Hylaeochampsa vectiana from the wealden of the Isle of Wight. American Museum novitates; no. 3032.
• Salisbury, S. W. & Naish, D. (2011). Crocodilians. In Batten, D. J. (ed.) English Wealden Fossils. The Palaeontological Association (London), pp. 305-369.
• Salisbury, S. W. & Frey, E. 2000. A biomechanical transformation model for the evolution of semi-spheroidal articulations between adjoining vertebral bodies in crocodilians. In Grigg, G. C., Seebacher, F. & Franklin, C. E. (eds) Crocodilian Biology and Evolution. Surry Beatty & Sons (Chipping Norton, Aus.), pp. 85-134.
• Sweetman, S.C., Pedreira-Segade, U., & Vidovic, S. (2014) A new bernissartiid crocodyliform from the Lower Cretaceous Wessex Formation (Wealden Group, Barremian) of the Isle of Wight, southern England. Acta Palaeontologica Polonica (in press)