November 13 , 2016:

by Mark Witton

Palaeoartistry is not a science. Even a conservative reconstruction requires artists to stretch data and evidence further than would be allowed for any scientific study and the whole process relies more on inference and speculation than many of us would like to admit. Yes, palaeoart is data-led and evidence-based, but only in rare circumstances do we have enough data to bring us to a single, reliable interpretation of a fossil species. Most of us would agree that there are some aspects of reconstructions that we can and should be getting 'right' for many species - basic proportions and musculature being top of the list - but beyond these science can often only narrow our choices, not present definitive answers. In lieu of clear scientific guidance, what guides these decisions may be our personal preferences, logical thinking, the demands of a project, or the penchants of our consultants. This means that, odd as it may seem, vastly contrasting reconstructions can be construed equally credible. A weird, alternative take on a fossil species might be just as ‘accurate’ to our knowledge as another preferred or familiar one. When evidence is equivocal for two or more states, we have to concede that one interpretation can be just as 'correct' as another.

With this in mind, I thought it might be pertinent to talk about the above reworked reconstruction of Plateosaurus engelhardti with a filamentous coat, and why – at time of writing at least - it’s perhaps neither ‘right’ or ‘wrong’ to depict this animal in this way. I could have covered it with scales as an alternative and not necessarily been ‘correct’ or ‘incorrect’, either. The question of whether sauropodomorphs were wholly scaly has not escaped discussion in many quarters - it was even mentioned at this blog briefly back in 2013 - so, for a change, and perhaps to mirror the sometimes antagonistic way that similar matters are discussed on social media, I’m going to present my thoughts here as a conversational debate between… myself. The idea is that it will allow for fuller discussion of opposing points, but I suspect it really just reflects the amount of time I spend alone at home with no-one but some chickens and various squamates for company. Whatever, hopefully the 'conversation' below will be balanced: both ‘pro-filaments’ and ‘pro-scales’ have important points to make, and I’m not strongly advocating one or the other here: the point is that both sides have valid points to make, so warrant an equal platform. Over to, er… me, then.

Me¹, meet Me²
Me¹ (filaments, opening statement). Should we restore Plateosaurus and other sauropodomorphs with filaments? Maybe. The evolution of dinosaur integument is an increasingly complicated area of study, and the idea that scales alone were the most likely ancestral condition for the major dinosaur lineages is no longer certain. As is well known to many, in the last decade we’re discovered filaments occurring in not only theropods but also in disparate parts Ornithischia too, and detailed new studies are suggesting that filaments in the likes of Psittacosaurus are structurally similar to those of other ornithischians as well as extant and extinct theropods (Mayr et al. 2016). They may even be similar enough to suggest true homology (Mayr et al. 2016), which strongly implies dinosaur skin may have been at least partly fuzzy in its ancestral form. Indeed, with pterosaurs thrown into the mix as well, an ancestrally-filamentous Ornithodira remains not only a valid hypothesis, but one that has passed several important tests in recent years. This being so, a filamented Triassic sauropodomorph is a sensible extrapolation of modern data.

Me² (scales, opening statement). Should we restore Plateosaurus and other sauropodomorphs with filaments? Maybe not. Firstly, studies have shown that our reconstructions of the ancestral ornithodiran integument type remains highly sensitive to the condition of its basalmost species, and we lack fossil data on these forms (Barrett et al. 2015). The 'ancestrally filamentous hypothesis' is enjoying some invigoration from new discoveries and research, but the game is not over yet. Virtually all of our dinosaur skin samples stem from derived species that had plenty of time to modify their integument from the primitive condition, and we have to concede that - whatever we think about the ancestry of dinosaur skin - they were very plastic in integument types. Thus, an important test of this hypothesis will be the recovery of good fossil skin samples from Triassic dinosauromorphs and pterosaurs, and their close relatives. Until we find these, or a fuzzy sauropod fossil, the recovery of scales from all three major dinosaur clades means the argument for 'ancestrally scaled' remains valid.

Me¹. Three points in response here. Firstly, admittedly playing Devil’s Advocate, a <10% chance of sauropodomorphs being filamented is still a chance, right? A filamentous Plateosaurus may not show what is most probable, but it still shows something that science shows is ‘possible’.

Secondly, and more constructively, the fact that skin impressions are not known outside of true sauropods means we may want to question what that the sauropodomorph stats of the Barrett et al. (2015) study really tell us. Does it reflect the condition for all sauropodomorphs, or just Sauropoda? The same probability assessments gives a 50% chance of filaments being ancestral to Saurischia, so the the first sauropodomorphs must have a somewhat higher chance of being filamentous, or at least being closely related to filamented species. Presumably, that 50% chance of filaments doesn’t just plummet the moment we steer evolution to the sauropod line: it’s a long evolutionary road from a basal saurischian to the sorts of sauropods we have with skin impressions, and we have no idea if or when filaments were abandoned on that road. We have a data vacuum of skin at the base of Saurischia: after sauropods, the next closest saurischian with skin impressions to Plateosaurus is the abelisaurid theropod Carnotaurus - hardly a close relative at all. Our absence of skin impressions around the phylogenetic neighbourhood of Plateosaurus, and our data about the likelihood of filaments in saurischians as a whole favours open-mindedness about the life appearance of these animals.

The third point is that if recent claims about dinosaur filament homology are correct, we have to assume that these structures were present in some form in the stock that gave rise to all major clades. Seeing as theropods retained filaments after the theropod/sauropodomorph split at the base of Saurischia, we should probably assume that sauropodomorphs lost their filaments after that divide. If so, a fuzzy Triassic sauropodomorph is not a far stretch.

Me². But - even assuming homology of filaments - if Carnotaurus is scaly, and so are sauropods, we can contrarily hypothesise that saurischians were secondarily-scaly ancestrally. This might even be the most objective reading of the data we have.

Me¹. Perhaps, but is the data supporting that interpretation really reliable? Carnotaurus is actually a weird outlier among theropods, it being the only theropod known with extensively scaly skin impressions. We have to wonder how significant this is against the wider backdrop of extensively filamented coelurosaurs sitting just a little higher up the theropod tree. As the rootward-bracket of the theropod integument bracket it's almost irritatingly important - it has a lot of sway in our reading of dinosaur integument evolution - but we still have to view it as a single outlier against the wider picture of theropod integument. As with any outlier, we have to be cautious about over-interpretation, or thinking one datum can give us the whole picture. As with so many palaeontological issues, we need more information.

And if we need an example of how sensitive our dataset still is, we need only consider Psittacosaurus, Kulindadromeus. Both are deeply nested within Ornithischia but basal to clades dominated by scaly species, and yet both have filaments. No-one would have predicted their integument type from their relatives. Not only does this show that our data may not be reliable enough yet to make confident predictions about integument types, but it suggests skin types might have been quite a bit more varied among even closely related dinosaurs than we anticipated.

Me². The risk here is that we’re pandering to exceptions, unknown data and slim chances. Arguments about the unknown nature of sauropodomorph or early saurischian skin seem like threading loopholes more than effective rebuttals. They play on what we don’t know rather than what we do, and that’s not how science works. There’s lots to be said for keeping an open mind, but we shouldn’t ignore data. Sure, there’s room for doubt here and we may be proved wrong in the future, but palaeoart should probably err on the side of caution, using the best supported, highest probability models to inform reconstructions. ‘Being wrong for the right reasons’ is perhaps the motto we should take when faced with the data gulfs associated with restoring partly known animals.

Me¹. The flip side of this is that ornithodiran integument has been proved complicated and surprising often enough that assuming variation in the poorly known areas is justified. Who expected Kulindadromeus and Psittacosaurus, or Tianyulong? Who, for that matter, would have predicted the first fluffy pterosaur fossils among - at that point - entirely scaly relatives? The point about exploiting unknown data is an important one, but we have a strong precedent for filaments in poorly sampled areas of ornithodiran evolution now. This is less exploiting a loophole than admitting we don’t have a full picture yet, and simply portraying one of the two more likely options of integument form.

Furthermore, Kulindadromeus and Psittacosaurus are great examples of how dangerous our approach to integument reconstruction is when we only have scraps of soft-tissue. It’s only because of their extensive soft-tissue preservation that we know they mixed scales and filaments in different body regions. And it’s not just these dinosaurs that show us that. Pterosaurs have scaly feet to counter their fluffy bodies (Frey et al. 2003), and the extinct mammal Spinolestes is known to have had scales, bristles, and variably long and short fur (Martin et al. 2015). Andrea Cau has even cast doubt on our presumed reasonable knowledge of Carnotaurus skin, pointing out that its skin impressions all pertain to the underside of the animal and that the dorsal surface could be entirely different. We thus have to ask what we really know about sauropod skin: are the bits we have representative of whole animals, or the group as a whole? The most extensive set we have - those from a diplodocid that might be Kaatedocus, described by Czerkas (1992) - show a lot scaling on the body, which meets the predictions we’ve made from smaller pieces of skin found with other sauropods. But it might be naive to think this offers a significant insight into these species, or rules out the chance of localised filaments on some sauropodomorph species.

Me². But where do we draw the line here? There has to be a point where we can say ‘we haven't seen evidence of filaments yet, and we should factor this into our science’ without someone going ‘you don’t know the whole animal yet!’. Some artists take this to an extreme, restoring animals like Edmontosaurus with large filamented regions despite this species being known from several well-studied and extensively-scaled mummified individuals. These have no evidence of filaments whatsoever, despite preserving scales down to millimetre resolution, and yet some folks are still unconvinced, speculating that filaments were poking through gaps between scales and so on. Palaeoart like this Plateosaurus reconstruction almost holds palaeontology to a standard of knowledge that it’s unlikely to ever attain: no, we don’t have skin impressions from every species, we don’t have good skin impressions from many species at all and fossils are never perfect records of animal appearance. But we have to use what we have: science does not work on a philosophy of 'assume whatever until proven otherwise'.

Me². It must be said that it would be easy to construct this conversation around a scaled version of this animal, and discuss why it doesn’t have filaments. Our base expectation for dinosaur integument and life appearance is in a state of flux, no matter what we personally prefer or assume.

Me¹. I think a point often lost on viewers of palaeoart is that these artworks are not, and cannot, be definitive, incontrovertible renditions of these animals. There are some animals so well represented in fossils that they lend themselves well to ultra-detailed reconstructions which are hard to quibble over to significant degrees - the awesome Bob Nicholls Psittacosaurus model being a great example (Vinther et al. 2016) - but for lesser known animals like Plateosaurus we are only painting hypotheses, not fact-based reality. This painting is one possible reconstruction of Plateosaurus as known in 2016, a time when interpretations of dinosaur skin evolution remain in flux. Time will tell if it's the product of over-interpretation of fossil data, or a lucky gambit later borne out with fossil evidence. I don’t mind getting stuff like this wrong: I’m more interested in painting and exploring credible possibilities of what we know now, not being ‘right’. We may never know what is ‘right’, so there’s not much point worrying about it. There are a couple of essays on this topic in my new book, Recreating an Age of Reptiles (Witton 2016).

Me². You’ve seen RecARep? I hear it’s awesome and that everyone should buy a copy.

Me¹. Wow, that’s desperate. Any actual final points?

Me². In a previous post on the role of pterosaurs in interpreting dinosaur filaments I concluded that: “Forcibly arguing for either scales or filaments at the base of Dinosauria seems premature at this stage, and, whatever our personal hunches are, it seems sensible to accept some ambiguity in this situation for now.” I think that’s true here too. There are certainly arguments to be had on both sides, some stronger than others, but neither side has knock-out data or evidence on the table yet. It’s the same old frustrating cop-out, but we need more fossils, and fossils of the right sort, to resolve this. Specifically, we need early saurischians or dinosauromorphs with good skin preservation, as well as that Triassic sauropodomorph with excellent skin remains. It must be said that these animals are not generally found in fossil Lagerstätten conducive to good soft-tissue preservation, so I’m not advising anyone to hold their breath for this one. But new techniques for detecting soft-tissues and increasing awareness of soft-tissue preservation in lithologies once thought to only preserve bone are reasons to be optimistic that we'll have insight on these matters one day.

Me¹. And ‘frustrating’ is the right word here, too. It seems like dinosaur science has made sufficient headway on understanding integument evolution and predictive methodologies that a reasonable, if provisional answer to the ancestral integument of the three major clades is close. But the puzzle piece needed to get our first good look at the broad picture is still out of reach.

Me¹. Pfft... Good… bad… I’m the guy with the graphics tablet.

Me². Movie quotes in scientific blog post don’t make you look clever, you know. You just cheapen the whole act.

Me¹. There is no fate but what we make for ourselves.

Me². What...? that doesn’t even fit our context.

Me¹. It wasn’t the airplanes. It was beauty killed the beast.

Me². Sigh, why do I hang around with you? I think we're done here.

So long everyone - I'm away from my computer for the next few weeks so I'm going to be pretty quiet in blog comments, social media and so on. Things will pick up again come December when we'll be addressing the sauropods in the palaeo-outreach room: has the popularity of dinosaurs above other fossil animals become a problem?

This blog post was inarguably supported by Patreon
The paintings and words featured here are sponsored by an excellent group of animals with regional variation in integument, 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. Later this month (much later - around the 28th/29th) I'll be uploading a video version of the presentation I gave over the Halloween weekend at Dinosaur Days 2016, entitled Palaeoart and the Never-Ending Quest for Accuracy. Here's the title slide to whet your appetite:

http://markwitton-com.blogspot.ru/2016/11/the-blog-post-where-i-ask-myself-should.html