Skin-deep: the 'One Skin Fits All' approach to integument reconstruction in palaeoart

The snowy, chilly plains of Maastrichtian Alaska, where Pachyrhinosaurus perotorum roamed. But were they scaly like other ceratopsids, or covered in protofeathers, as shown here?

So... no pressure here, then. I innocently replaced my festive Facebook profile picture with some detail from the painting above last Sunday (07/01/13) and quickly found a storm (well, gusty conditions) of discussion, 'likings' and shares, with several folks mentioning their anticipation of this post to see what all the fuss is about. The Facebook responses have been interesting and mixed: nods of approval, some head-slapping 'why didn't I think of that', a revelation that Tom Hopp already did this last year and, perhaps more predictably, scepticism from a number of individuals who consider the whole concept very silly indeed. In short, there seems to be a certain amount of expectation in the air about this image, and I wouldn't be surprised if some virtual beer bottles are hurled at me across the Internet should this explanation not prove convincing. Here's hoping I've done my homework properly, then.

The painting in question shows a family of the Alaskan centrosaurine Pachyrhinosaurus perotorum, a species notable for its existence in rather chilly, latest Cretaceous climates at palaeolatitudes of 80-90°. It differs from other pictures of this species by having its Muskox Quotient upped by 500%, replacing the scaly hides of more traditional Pachyrhinosaurus reconstructions with a blanket of fuzz analogous to the fuzzy, unkempt feathers of modern ratites. Fuzzy polar dinosaurs are not unusual in palaeoart nowadays and they result in animals that look immediately more at home in icy, subfreezing climates than their scaly brethren. This image, however, directly contradicts what most folks will say we know about horned dinosaur integument. Some comments on Facebook have already wheeled this argument out: known ceratopsian integuments were predominately scaly, so the concept of a shaggy pachyrhinosaur is nonsense, right? Well, I'm going to argue here that it's not, or at least not a concept that is easily dismissed. Before we go any further, it's worth stressing that I'm not presenting this image as the new 'standard' for Pachyrhinosaurus perotorum: I don't know of any new evidence that confirms the shaggy hides shown here, be it soft-tissue remains of ceratopsids or a new interpretation of dinosaur evolution that suggests super-fuzzy ornithischians were common. Nor, for that matter, do I have the heads up on research indicating that latest Cretaceous Alaskan palaeoclimates were much lower than expected. Instead, across four points, I'm going to argue that, based on what we know of dinosaur evolution, the responses of modern animals to their environments, and - importantly - the vast gulf of unknown data regarding dinosaur appearance, that this concept is as plausible as our scaly variants and, in some respects, may be more plausible. On the way, I'm going to suggest that, as with some other considerations in palaeoart, we may be too conservative when it comes to depicting animal integuments, because we focus too much on their evolutionary relationships without considering their likely adaptations to habitats and lifestyles. Hmm... this is all starting to sound very All Yesterdays, isn't it? That's not a coincidence.

1) Ornithischians were fuzzy, and some were probably fuzzier than others.
First up, the least controversial pin in this case. Thanks to Tianyulong and the early ceratopsian Psittacosaurus, we know that ornithischians were covered in more than just scales, the former being covered in filamentous structures akin to early feathers and the latter possessing long quills (Mayr et al. 2002; Zheng et al. 2009) Accordingly, it's now fairly fashionable, and by no means unreasonable, to restore even large ceratopsids with at least a smattering of quills across their bodies like those seen on Psittacosaurus, reflecting a relict integument from an earlier phase of their evolutionary history. It naturally follows that we should expect some taxa to have been more densely adorned with filaments and quills than others, just as fur and feathers are of variable densities in modern species. Accordingly, while a shaggy pachyrhinosaur is certainly at the 'extreme' end of our predictions for an ornithischian integument, it does not directly contradict anything we know regarding dinosaur evolution. Ceratopsids probably had the appropriate genetic blueprints to produce a shaggy animal, so long as the right conditions promoted its expression. There is a question of how appropriate it is to cover a ceratopsid in shaggy integumentary structures however, in light of preserved skin impressions of other ceratopsids. How likely is it that any horned dinosaurs were fuzzy?



Fossil integument of Chasmosaurus belli. From Sternberg 1925.

2) Is the extrapolation of preserved integuments to other species that reliable, really?
Skin impressions and the remains of other integumentary structures are Holy Grails to palaeoartists, and we use them extensively in restoring extinct animals. Through phylogenetic bracketing, or use of their basic phylogenetic proxy when less data is available, we stretch these remains over entire clades so that the known integument of one species becomes the norm for an entire group - what I'll call 'One Skin Fits All' approach. Thus, because we have scaly skin impressions for three ceratopsids - Centrosaurus, Triceratops and Chasmosaurus (see image, above, of the latter. From Sternberg 1925), it's assumed that scales were common to the entire clade. There's nothing necessarily wrong with this assessment and, one may argue, it's the most parsimonious way to interpret this data. A stick in the mud, however, is that another dataset, the diversity of integuments in modern animals, suggests that integuments can vary wildly within groups, and that we could be vastly underestimating the integumentary variation in extinct animals.

Consider the different varieties of fluff, fur, feathers, hair, bristles and other fuzzes in a group of modern animals and then think how a future palaeoartist would reconstruct all varieties of that group if they only had access to only one or two examples of integument. Perhaps all reconstructions of bovids would have woolly coats like those of sheep, or, conversely, the sparse, almost naked skin like a water buffalo? We may deck all primates out in the long capes of colobus monkeys, all pigs with boar-like fur, or cover every inch of birds with feathers. We know such approaches are wrong because these groups demonstrably show variation in the distribution, length and structure of their varying integuments, and yet we maintain a One Skin Fits All' approach to fossil clades. We can't even play the 'extremely close relationship card' in this game as the likes of woolly mammoths, and the fuzzy Sumatran and woolly rhinos, show vastly different integuments to their closest, naked relatives.

One could counter this point by arguing that the relative abundance of scaly remains in certain dinosaur lineages suggest that most, if not all members of that clan were scaly. Perhaps, but we should consider both the sample sizes here and the taphonomic window through which fossils are passed to the modern day. We have, at best, skin impressions from a handful of species compared to the group diversity, so statistical support for the 'One Skin Fits All' approach is low. Moreover, which types of skin are more likely to be preserved? Taphonomic observations on modern animals suggest that fur and feathers are easily removed from carcasses by biological or physical processes, so their preservation potential in ancient animals is low outside of fossil Lagerstätten. Is it a coincidence that the only skin impressions we find outside of Lagerstätten are scaly, leathery hides? I don't have the answer to that question, but it's worth chewing over.

With all this in mind, I wonder if applying the fossil integuments of one species to all its relatives, even close ones, is a questionable practise. I'm not saying that skin impressions are useless and that we should pay them no attention, but we should remember that they only highlight possibilities and perhaps some degree of probability for integumentary structures in a related species. They may well also have no bearing whatsoever on the appearance of their relatives. We're dealing with a great amount of unknown data when reconstructing ancient integuments, and we know how complex this issue is through modern species. When applying this thought to horned dinosaurs, we can say that the scaly skin impressions we have for a few species demonstrate that some bore scales, but we cannot rule out the possibility that others were covered in entirely different structures, like the quills and fuzz that seem deeply rooted in dinosaur ancestry. It does seem likely that many centrosaurines heads, including Pachyrhinosaurus, bore heavily keratinised scales and pads (Hieronymus et al. 2009), but, of course, this doesn't tell us much about the rest of the body. The majority of skin in Pachyrhinosaurus could be scaly, fuzzy, or anywhere inbetween. Without skin impressions to directly tell us the integument of specific species, there's no way to be sure. We need to be careful that we do not afflict ourselves with palaeoartistic phyloblindness here, by only considering these animals as denizens of cladograms and evolutionary hypotheses. Phylogenies may tell us what is possible for integument reconstructions, but other factors may help us decide is more probable.

3) Phylogeny is far from the only factor controlling integument, and can be readily overruled
It's funny to think that, for all the time we spend looking at the phylogenies of extinct animals, we're often missing much about the raw power that drove their evolution: adaptation. This is probably because we're lacking so much anatomical detail in their fossils that their responses to even broad environmental changes are largely undetectable, so understanding why they change through time is not always as certain as how. Nevertheless, we can be sure that different environments drove modification to the anatomy of extinct lineages on small and large scales, and integuments were likely to be one of the most affected tissues. Animal integuments are critical interfaces between body and environment, and have to be appropriately adapted for given habitats. This is a readily observable phenomenon in modern animals, because their integuments reflect all sorts of environmental factors including sun exposure, temperature, local vegetation types, water availability, parasite prevalence, and their local predators. Presumably, this is why such variation in integument exists in even closely related species. But we frequently reconstruct fossil species as if they all live in the same place. Sedimentological and isotope data reveals that closely related ancient species sometimes lived in starkly contrasting settings, but because we frequently take the 'One Skin Fits All' approach, our animals look very similar, irrespective of the requirements of their habitats.

Hot Fuzz: a reference to the condensing breath of the animal, the controversial concept depicted here, or just an excuse for a bad pun? Whatever: it's an excuse to link to this clip from Hot Fuzz: Best. Granny. Kick. Ever.

Behaviour may also have an effect on integument. Sun-shy, non-aggressive and cursorial animals may well bear thinner integuments than slower, frequently exposed or bad-tempered species, for instance. Morphology, too, will have an influence, with larger animals having lessened needs for insulation or being capable of carrying heavier, armoured hides. These are all things palaeoartists should be considering when reconstructing extinct animals: they should look adapted to the lifestyles we predict for them, rather than being based on phylogenetic hypotheses alone. Perhaps some desert-living dinosaurs had elephant-like, deeply wrinkled skin to help heat dissipation, while smaller desert-dwellers had extremely short feathers, or none at all, to prevent overheating. Maybe large theropods had heavy scales on their faces to defend themselves during bouts of head biting. We don't know for sure, but we can be certain that these species had to be appropriately adapted for wherever and however they lived. In short, we need to be wary reconstructing our ancient species by cladogram alone, and realise that integuments, and soft-tissue anatomies in general, should reflect a combination of phylogenetic data and possible adaptations to habitats and lifestyles. This, undoubtedly, involves some of the 'informed speculation' that has been discussed so much with All Yesterdays, but the results are more consistent with our knowledge of modern animals and evolution generally, and arguably producing a more convincing look into the ancient world ('convincing' is the right word here: I'm not sure we're ever going to get 'accurate' results in this game). In fact, after mulling this over for some time, I find the typical and conservative, 'One Skin Fits All' approach much harder to defend than the more open minded, environmentally-influenced reconstructions argued for here.

Bringing this back to our fuzzy pachyrhinosaurs, we again have to question the how applicable the currently available ceratopsid skin impressions are to this Alaskan species. The scaly hides of Chasmosaurus, Centrosaurus and Triceratops represent animals living in more southerly regions than Pachyrhinosaurus, which were at least temperate to subtropical in climate. The former two taxa also lived somewhat before Pachyrhinosaurus, when global temperatures were, on average, a little warmer. These sub-arctic coastal plains encountered by Pachyrhinosaurus perotorum, by contrast, were much cooler, and sometimes genuinely cold. Accordingly, the selection pressures on integument may have been very different for P. perotorum compared to these warm-climate ceratopsids, and we have to wonder how suitable the skin impressions of Chasmosaurus et al. are for reconstructions of Pachyrhinosaurus. We wouldn't, after all, expect the integument of a yak to resemble that of a African buffalo, or consider the fur of a lithe gazelle a suitable model for mountain goat fur. With this philosophy in mind, the question of shaggy pachyrhinosaurs shifts focus from arguments about the cladograms and the skin impressions of their relatives, and on to whether or not Late Cretaceous Alaska was cold enough to promote the development of an extreme integument adaptation in a large dinosaur species.

4) Late Cretaceous Alaska: a struggle for any tourist board
Although nowhere near as bleak as our modern Alaska, the dinosaur faunas inhabiting the northern reaches of latest Cretaceous Alaska would have experienced fairly grim weather for much of the time, perhaps akin to that experienced by modern animals living on the northwest coast of Canada or the more depressing parts of Scotland. In a recent review based on palaeobotanical data from the Cretaceous Arctic, Spicer and Herman (2010) suggested that uppermost Cretaceous Alaska experienced mean average temperatures around 6°C, with summer months attaining a comfortable 14.5°C, but winter months dropping to an average of -2°C. The Pachyrhinosaurus perotorum-bearing Prince Creek Formation may have been a little cooler than other parts of the Late Cretaceous arctic circle, with a mean average temperature between 2.5-5°C. Winter lasts a long time at 80-90° latitude, with 5 of darkness bracketed by 2 months of twilight. A permanent cloud cap over the Late Cretaceous Arctic (detected by the oversize nature of the fossil plant leaves from Cretaceous Alaskan localities) acted as a atmospheric blanket for the region, prohibiting temperatures from plummeting below freezing low for long period. The lowest temperatures - perhaps -10°C - may not have lasted more than a few weeks. Evidence for deep freezes is absent however, with neither the sedimentological or palaeobotanical record indicating nothing more severe than week-long frosts and light freezes. Despite this, rain and snow were probably common, with relative humidity averaging about 80% and even the driest months of the year experiencing over 180 mm of rain. The collective three wettest months, by contrast, collected almost 800 mm. (To put this in perspective, rain-soaked England has an average annual rainfall of 854 mm, according to the UK Met Office [via Wikipedia]). Such a climate was capable of supporting a rich array of plantlife, and evergreen taiga-like forests were common, as were swamps, rivers and other bodies of water. Such conditions seem fairly common right the way through Late Cretaceous Alaska, with conditions a full 8° south of the Prince Creek Formation seeming similarly cold and wet. I should add that this consideration of ancient Alaska  isn't particularly controversial, the palaeobotanical data mentioned here matching palaeoclimate models based on isotope records, animal distribution and sedimentology.

So the Prince Creek palaeoenviroment wasn't exactly an ideal holiday spot, but was it 'extreme' enough to promote the evolution of an fuzzy coat in a 1.5-2 tonne dinosaur species? Given the dense furs we see in large mammals found in similar climates, I think it's certainly a possibility. It would certainly be far weirder if polar dinosaurs of the Late Cretaceous didn't respond to their climate somehow, and a thick coat of protofeathers is one possible adaptation to their cool, wet habitat. A layer of insulating fat would be another (cue an image of some tubby pachyrhinosaurs). Of course, the picture may be different if these animals hibernated or migrated in and out of Alaska annually, enjoying the brief mild period before heading south to escape the winter. The latter is perhaps the most widely discussed concept, but direct evidence for such migrations in Alaskan dinosaurs is as sparse as evidence for their fluffy integuments or fat layers. In the concept proposed here, the protofeather coat may have acted as an insulator against the cold, prevented wind, rain and snow from hitting the naked skin of the animal, or both. The latter function would benefit from the coat being thick and fluffy, but this may not have lead to overheating even in a big animal like Pachyrhinosaurus: the ragged, loose coats of hot-climate adapted ratites appear similarly thick and massive without overheating their owners. Perhaps the concept of shaggy coats in these Alaskan dinosaurs doesn't seem so as implausible as it may first appear, then.

To bring this more-mammoth-than-intended essay to a close, then, I again stress that I'm not saying 'this is what Pachyrhinosaurus looked like!', but attempting to present it as a product of both its phylogenetic history, its environment and habits, and not simply reconstructed via a cladogram. I think there's a lot of scope for these sorts of palaeoartistic renditions, even it does mean more reliance on the 'informed speculation' principle of All Yesterdays. Of course, this whole argument flows back to the core idea behind the All Yesterdays movement: the conservative, 'One Skin Fits All' approach to integument reconstruction is just as likely to be wrong as our more speculative concepts, but at least the use of informed speculation lines the reconstruction up with our knowledge of modern animal diversity.

References

• Hieronymus, T. L., Witmer, L. M., Tanke, D. H. and Currie, P. J. 2009. The facial integument of centrosaurine ceratopsids: morphological and histological correlates of novel skin structures. Anatomical Record, 292, 1370–1396.
• Mayr, G., Peters, D. S., Plodowski, G. and Vogel, O. 2002. Bristle-like integumentary structures at the tail of the horned dinosaur Psittacosaurus. Naturwissenschaften 89, 361–365.
• Spicer, R. A. and Herman, A. B. 2010. The Late Cretaceous environment of the Arctic: A quantitative reassessment based on plant fossils. Palaeogeography, Palaeoclimatology, Palaeoecology, 295, 423–442.
• Sternberg, C. 1925. Integument of Chasmosaurus belli. The Canadian Field-Naturalist, 39, 108-1 10.
• Zheng, X., You, H., Xu, X. and Dong, Z. 2009. An Early Cretaceous heterodontosaurid dinosaur with filamentous integumentary structures. Nature, 458, 333–336.