The return of the scaly T. rex to modern paleo-art

Facebooktwitterredditpinterestlinkedintumblrmailby feather
Tyrannosaurus rex walking towards camera. Art by John Sibbick.

[NOTE: Post has been updated to include a section on scale size]

This has certainly been an interesting year. Two papers dropped in the past three months that have put the brakes on a recent trend in paleo-art. That trend? Why the feather-coated T. rex of course.

First, in March, we saw the release of a paper detailing a new species of Daspletosaurus and its relationship to D. torosus.

Carr, T.D., Varricchio, D.J., Sedlmayr, J.C., Roberts, E.M., Moore, J.R. 2017. New Tyrannosaur with Evidence for Anagenesis and Crocodile-Like Facial Sensory System. Scientific Reports. 7(44942):1–11.

In this paper, Carr et al. argue for the designation of a new Daspletosaurus species, D. horneri. The authors argue, based on skull shape and chronostratigraphic position, that D. horneri was the direct ancestor to D. torosus. I thought that the authors put forth a compelling argument for this anagenic event and backed up their position well. Interestingly, this part of the paper should have been the most controversial. As anyone who has read anything from Horner and Scanella over the past eight years can attest, arguing for a direct ancestor-descendant relationship for dinosaurs is difficult to do and even harder to win over others in the field. So it is somewhat surprising to see a case for anagenesis in Daspletosaurus taken so well by the palontological community. All the more so given that it involves a tyrannosaur, the poster children for “cool guy” dinosaurs.

Instead, the most controversial part of the paper wound up being their soft-tissue reconstruction of the face for D. horneri. The author responsible for the soft-tissue reconstruction was Jayc Sedlmayr of Louisiana State University. Sedlmayr did his doctorate on osteological correlates for vasculature in extant archosaurs (birds & crocs). He is the seminal alumnus of the WitmerLab and thus is well within his wheelhouse for this type of soft-tissue reconstruction. Sedlmayr borrowed heavily from the work of another WitmerLab alumnus, Tobin Hieronymus, whose PhD work involved osteological correlates for integument on the skulls of animals. Although the skin is often well away from the underlying bones on most of the body, there are exceptions when it comes to the skull. There, areas that are not heavily muscled, tend to show intimate connections between the skin and the underlying bone. Hieronymus used these connections to determine how different integumentary appendages (scales, hair, feathers) affect the underlying bone (Hieronymus & Witmer 2007; Hieronymus et al. 2009). The authors found that the surface texture along the skull of D. horneri was “hummocky”. That is, it was covered in lots of closely packed ridges. According to Hieronymus & Witmer (2007), this texture correlates to scales as the overlying integumentary appendage. Thus, according to the authors, D. horneri had a scaly face (this is grossly oversimplified as the authors were able to piece together a variety of different integument variants along the skull, but you get the idea).

Scaly tyrannosaur cannonball one had been shot.

Then two weeks ago, we saw the release of another paper on tyrannosaur integument. However, unlike the previous paper, this one was specifically dedicated to integumentary types in tyrannosaurids.

Bell, P.R., Campione, N.E., Persons, W.S., Currie, P.J., Larson, P.L., Tanke, D.H., Bakker, R.T. 2017. Tyrannosauroid Integument Reveals Conflicting Patterns of Gigantism and Feather Evolution. Biology Letters. 13:20170092

In this paper, the authors set out to survey all known instances of “skin” impressions for tyrannosaurids. Their list of taxa included Albertosaurus, Tarbosaurus, Daspletosaurus, and Gorgosaurus. Their results pretty definitively indicated that scales were the predominant integumentary appendage on tyrannosaurids. The authors then went on to speculate why that would be if earlier tyrannosauroids had filamentous integument. They performed an ancestral character state reconstruction based on Parsimony and Bayesian-based trees from Brussatte and Carr 2016. Their results found that filaments came out strongly as the ancestral character for tyrannosauroids, but by no later than Tyrannosauridae proper, a reversion to scales had taken effect. The authors attributed this to body size evolution. Namely, larger tyrannosauroids reverted to scales over protofeathers.

Cannonball number 2 had just been shot.

The new orthodoxy

The cover photo from Nat Geo’s infamous Archaeoraptor announcement. Art by Mike Skrepknick

The response to these two papers have been interesting in light of a very noticeable trend in paleoart over the past thirteen years; the feathered Tyrannosaurus.

In truth, feather-covered Tyrannosaurus depictions have been around for even longer than that. Most notably, National Geographic featured a down-coated baby Tyrannosaurus rex on the cover of their ill-fated 1999 issue (featuring the infamous Archaeoraptor fiasco). However, this depiction of T. rex,or tyrannosaurs in general was few and far between back then. It really wasn’t until the discovery of Dilong paradoxus in 2004, and its phylogenetic placement as basal to tyrannosaurids, that the pendulum start to shift radically towards the enfluffenned T. rexD. paradoxus was a small theropod, though, so it could still be argued that large tyrannosaurids would have had a sparse feathering, if any at all. This argument stemmed mostly from comparisons to elephants, rhinoceros, and hippopotamus.

However, even this argument fell by the wayside when in 2014, Xing Xu and colleagues announced the discovery of a giant filamented theropod that nested close, but still basal to, Tyrannosauridae. The animal in question was, of course, Yutyrannus huali, an animal that I have written about at great length before. At 1400 kg, Y. huali was not a small animal, and yet it was adorned from head to toe with long, shaggy filaments (likely stage 1 protofeathers). If an Allosaurus sized theropod could still sport plumage, then the argument for other tyrannosaurids being fluffy suddenly appeared plausible.

And so the rampant enfluffenning began.

A cursory Google image search for T. rex, using time limiters for 2004–2008, 2008–2011, and 2012–2017, shows an exponential increase in the amount of enfluffenned T. rex over time. D. paradoxus planted the seed, but Y. hauli set off the bomb.




Tyrannosaurus(or near enough) art through indicated spans of time. Coloured images indicate enfluffenned animals. Totals are in parentheses. Search was limited to Deviant Art (and only one page of search results) to keep hits relevant. Click to enlarge

Now, admittedly, the image above is limited to Deviant Art images only. This was done both to keep the hits relevant (there’s lots of crap out there that has Tyrannosaurus associated with it), but also because of the vibrant paleo-art community on that site. Many dinosaurphiles frequent the website and many aspiring paleo-artists use DA as a portfolio for their work.

Given the fervor with which this trend has caught on in paleo-art, one could imagine that the latest two papers did not go over very well. Both papers have been greeted with intense skepticism. As a scientist, I can respect healthy skepticism about a new study (more studies should be treated this way), but I also expect the skeptics to be skeptical for the right reasons (i.e., the science). In the case of both Carr et al. 2017 and Bell et al. 2017, I have not really seen this play out.

I’m getting ahead of myself here. So let’s take a look at each new paper on its own.

Tyrannosaurid heads: scaly or just cracked?

Head scale reconstruction in D. horneri (left) vs. pseudoscales on head of crocodile. (right). D. horneri image by Dino Pulerà. Croc image from Milinkovitch et al. 2013

The largest complaint leveled at the Carr et al. paper was their use of osteological correlates for scales. The reason was that the authors compared the hummocky texture on tyrannosaurid skulls with that of crocodylians. This was done, in large part, because crocs are the closest scaly relatives of dinosaurs. So from an extant phylogenetic bracket perspective, it makes sense.

The problem, however, is that crocs don’t have scales on their heads.

Croc head “scales” are like fingerprints. No two are exactly alike. Image from Fig. 1C of Milinkovitch et al. 2013.

Some of you may be crying foul on this, but it’s true. Despite a superficially scaly appearance, crocodylians have recently been found to not have scales on their faces. Work by Michel Milinkovitch and others (2013) discovered that the “scales” seen in crocodylian heads, do not correspond to true, epidermal scales. True scales show serial homology. That is, each scale is a nearly identical unit to its neighbours. There are still variations, but the variations occur across multiple scales, rather than each individual scale. Scales are also consistent between members of the same species (for the most part). This consistency is extremely important for herpetologists, as many lizard and snakes are described by their facial scales. Crocodylians, on the other hand, do not show this consistency. Each crocodylian individual has a unique “scale” pattern on its face, and the pattern is randomly distributed so that no two “scales” look alike (well, barring coincidence). When the authors delved into the development of Nile crocodiles (Crocodylus niloticus), they found that the head gets covered in a single keratinous sheath that later cracks due to shrinkage, forming the random, scale-like patterns we see in modern crocodylians (Milinkovitch et al. 2013). The authors argued that this was likely due to the presence of dome pressure receptors (DPRs) along the face of modern crocodylians. These integumentary sense organs form before the corresponding integument and appear to have corrupted the scale-development process, resulting in the strange, pseudoscales seen on extant croc faces.

The implications of Milinkovitch et al.’s work is pretty damning, as it essentially means that neither extant members of Archosauria have scales on their heads. Does that mean that Daspletosaurus horneri and by extension, all other tyrannosaurids were also covered in a cracked, keratinous sheath?

Eh, probably not.

The reason why not comes from the next branch down from Archosauria, Lepidosauria. From Hieronymus et al. 2009 (emphasis mine):

A second form of osteological correlate for scales can be seen in many iguanian lizards, and consists of a regularly arranged, shallow, hummocky rugosity on the bone surface… These features are not related to intradermal ossification and are instead derived entirely from apophyseal bone growth beneath individual scales…

…both of these morphologies [the other one being osteoderms] are considered robust osteological correlates for epidermal scales in extinct taxa.

Reading through their paper, it was apparent that Hieronymus et al. found the hummocky texture on the skull bones of iguanian lizards, not crocodylians (though these were also studied). Thus, the osteological correlate for scales on top of bones comes from data in lizards, and not crocs. Lizard head scales are true scales, so the hummocky texture morphology still holds true. That the hummocky texture was said to be associated with croc epidermis can be blamed on a misread of Hieronymus et al. 2009.

Rugosity profiles used to correlate soft tissues with underlying bone. Image from Fig. 4 of Hieronymus et al. 2009.

All that said, it is worth bringing up the fact that Carr et al. also argued for the presence of DPRs in tyrannosaurids based on similar neurovascular foramina density to extant crocodylians. This is not the first time that a theropod has been hypothesized to have had a “sensitive” face. Ibrahim et al. (2014) argued for similar facial sensitivity in Spinosaurus aegyptiacus, as have Barker et al. 2017, in their recent paper on Neovenator salerii. If true, these data would suggest that highly sensitive snouts were a common theme in theropod evolution, and that a cornified, but not scaly, face may have followed suit.

The problem is that this is almost certainly not the case.

The DPRs in crocodylians work by sensing pressure changes in the surrounding water. DPRs are the tetrapod equivalent to the lateral line system in fish. Both the lateral line system and DPRs work well underwater because water is an incompressible substance. Thus, pressure waves travel through water as if it were a solid. This is why the speed of sound is over four times faster in water vs. air. Air is comprised of compressible gases that make it harder for pressure waves to propagate very far. The inefficiency of sound travel in air is likely why tetrapods dumped the lateral line system in favour of ears. This means that DPRs are essentially useless out of the water. Any pressure changes that they could detect (barring explosions) would be at distances that are practically on top of the animal’s mouth. Hardly that helpful.

Trigeminal nerve space compared in Tyrannosaurus rex (top) and Alligator mississippiensis (bottom). Note the much chunkier trigeminal in the latter.
Top image adapted from Witmer & Ridgely 2009. Bottom image adapted from George & Holliday 2013.

Another strike against the super-sensitive snouts of theropods, comes from their brain endocasts. Crocodylian brains show a honking big foramen for the trigeminal nerve (George & Holliday 2013). Trigeminal is the main sensory nerve to the face of tetrapods. A larger trigeminal foramen indicates a larger trigeminal nerve, which means more available neurons to carry sensory information back to the brain. Compared to crocs, the trigeminal foramen in theropods is small. It’s still big (it’s the largest cranial nerve for most tetrapods), but it’s not that big compared to an animal of similar size (e.g., see Witmer & Ridgely 2009). An average-sized trigeminal foramen strongly argues against the super-sensitive snout hypothesis.

So if theropods didn’t have DPRs on their snouts then why do they seem to be so heavily innervated?

The answer lies in the suffix of the term: neurovascular (i.e., nerves and blood vessels). The foramina described by Ibrahim et al. 2014, Carr et al. 2017, and Barker et al. 2017, are known for carrying nerves and associated blood vessels. All the theropods that have been hypothesized to have had “super-sensitive” snouts, were multi-tonne creatures. Big animals have more cells than smaller animals, and those cells need nutrients. What we are likely seeing is the inevitable increase in vascular irrigation to the heads of large theropods. More blood vessels require either larger foramina to travel in, or more foramina to accommodate the extensive branching patterns of the arteries to their associated capillary beds. For big theropods, we see a mix of both. A good way to test this hypothesis would be to take similarly sized theropods and crocodylians and compare their respective foramina densities. I suspect that the crocs will beat out the theropods by a country mile.

Returning to the crux of this post, the complaints that I read online focused heavily on the comparison to crocodylians, despite the fact that the authors used osteological correlates for lizards. I’ve also seen comparisons to other gnarly-textured animals such as hippos, but these types of superficial comparisons are not equvalent. Carr et al. looked a minute details on the skull of D. horneri, and they painstakingly detailed not only the textures seen on the skull bones, but also what layers of textures were present, and the associated profiles those textures corresponded to. This is a far cry from simply looking at a roughened hippo skull and saying it is equivalent to a crocodile.

New tyrannosaur scale paper for old tyrannosaur scale imprints

Turning now to our second paper, the headlines that came out from this study were certainly entertaining. They ranged from groundbreaking find to full-on snark. I found the response from the paleophile community very interesting. In many ways it mirrorred the reactions seen two years ago when Paul Barrett and colleagues released their work testing the homology and ancestral integumentary appendage type of dinosaurs (Barrett et al. 2015). In the case of Barrett et al., the reaction started off quiet followed a few weeks later by blog posts and Facebook rants that took the paper to task over allegedly bad interpretations of the data. The Barrett et al. paper is worth reviewing, and is something I intend to do at a later date. For now, though it remains the only attempt to asssess homology in the filaments found in some dinosaurs and pterosaurs.

The case for Bell et al. 2017 was somewhat similar, except that the refractory period between quiescence and attack was much shorter. Some responses seem more a reaction to how the new articles were written, whereas as others were more even-toned in their critique of the paper.

In general, the complaints against the paper from Bell et al. can be summed up in one of three categories.

  1. The paper details “skin” impressions that have been known about for decades
  2. These impressions look no different than a rhinoceros ass / plucked chicken / dirt
  3. The authors are making blanket statements based off coin-sized samples of “skin”

Let’s tackle each complaint separately.

(1) The paper details “skin” impressions that have been known for decades

This complaint is a mix of fact and fiction. While it is true that Bell et al. looked at integument samples that had been previously collected, many of these samples had only been known by those in the paleo field (i.e., there was never a published account). The exceptions to this were the Wyrex “skin” impressions which were published in a book chapter by Neil Larson,  the Tarbosaurus impressions (Currie et al. 2003), and the Daspletosaurus impressions (Currie & Koppelhus 2015). So, yes, in some ways this is old news.

Here’s the kicker, though, most of these published impressions were mentions, not descriptions.

Let’s just look at these grand descriptions real quick.

The Wyrex specimen, which has been known about since 2004, was in privately owned hands for much of that time. It has only recently come to be housed at the Houston Museum of Natural Science, where it received an official accession number: HMNS 2006.1743.01. So despite knowing about this specimen for over ten years, it has only been available for scientific study in the last three years. The only published description of the material came from a book chapter in the 2008 book, Tyrannosaurus rex, the Tyrant King. Book chapters rarely get peer reviewed and are considered “grey literature” for scienfic publications.

So what did this little bit of grey literature say about the skin impressions?

During preparation, several patches of skin (Fig. 1.26B) were found with the skeleton. Most of the skin patches (more than a dozen) were found on the bottom side of the articulated tail. The discovery of skin with Wyrex is a first for T. rex. Plans are currently underway for futher excavation at the site in hope of finding more of the skeleton. — Tyrannosaurus rex, The Tyrant King, p: 47.

That’s it. A quick little blurb followed by the now infamous scale photo, is all we had for Wyrex. In contrast, Bell et al. spent an extensive amount of time detailing the various impressions found associated with Wyrex, and where they are located on the body. In other words, they actually described the impressions.

Okay, so what about the other “old news” specimens?

Here’s what Currie et al. (2003) had to say about other tyrannosaurid integument impressions:

Skin impressions (MPD 107/6A, Fig. 5D) were recovered from a Tarbosaurus skeleton destroyed by poachers at
Bugiin Tsav (N43° 52.164′
, E100° 00.605′). In this specimen, which is a large individual with a frontal width at the interorbital slot of 81 mm, the scales have an average diameter of 2.4 mm. The skin impression was recovered from the thoracic region of the body, although the damage done to the specimen makes it impossible to know exactly which region it covered.Other tyrannosaurid (Albertosaurus, Daspletosaurus, Gorgosaurus) skin impressions have been recovered from Alberta and Montana, and show the same lightly pebbled surfaces.  — Currie et al. 2003 p: 6

In one quick paragraph, Currie et al. briefly describe the scale impression found in Tarbosaurus, and say that it looks similar to multiple, undescribed scale impressions from three other tyrannosaurids. Prior to the landmark work of Bell et al., this was the best description of tyrannosaurid integument ever published.

Lastly, we have a “description” of scale impressionsin Daspletosaurus, by Currie and Koppelhus 2015:

Skin impressions associated with tyrannosaurid bones were first found when the holotype of Gorgosaurus libratus (CMN 2120) was being prepared further in the 1980s in Ottawa. Subsequently, skin (TMP 1994.186.0001) was identified with an Albertosaurus from the Horseshoe Canyon Formation of Edmonton. Traces of skin were found in the field with a large skeleton of Daspletosaurus (TMP 2001.036.0001, Fig. 2), and looked remarkably like that of Tarbosaurus from Mongolia (Currie et al. 2003). Unfortunately, although the skin impressions were collected, recent attempts to find them in the TMP collections were unsuccessful, suggesting they may still be in an unprepared block of matrix. Similar skin impressions (TMP 2003.045.0088) were found in the Albertosaurus bonebed (TMP locality L2204) and are currently under study. — Currie and Koppelhus 2015 p: 621

Once again, we learn of multiple integument impressions for different tyrannosaurids, all mentioned in a cursory examination of the material. No real description given, just a half-hearted mention of similarity to scale impressions that were briefly described in a paper from twelve years earlier.

Of the five species described by Bell et al. in their paper, three are being described for the first time, and all are being thoroughly described for the first time.

(2) These impressions look no different than a rhinoceros ass / plucked chicken / dirt

This argument tends to crop up a lot whenever someone mentions scale impressions in dinosaurs. Often not far behind is this ever so photogenic shot from the rear end of a rhinoceros.

Rhinoceros skin shown in closeup. Note the vaguely scaly appearance. Left image by Sanjay Ach. Right image by derausdo

Opponents to the scaly dinosaur idea seem to be real fond of rhino skin, and I can understand why. On the outset, rhino skin has a superficial resemblance to scales. However, if one does a side by side comparison between rhino skin and scales, the differences becomes very noticeable.

As I mentioned earlier, scales have serial homology. That is, each scale looks the same as its neighbours (within reason). That means one should be able to pick out a consistent pattern in scales that won’t be present in cracked skin. So compare the above to these shots from scaly animals:

Closeup of scales in an anole (left), crocodile (center) and Tokay Gecko (right). Animals chosen for their non-imbricating scales, most similar to what we see in dinosaurs. Croc scales by Leah Borchert

Compare the above image, with the rhino skin. Note how one can easily see the patterns in the scales (with patterns on top of patterns in some instances), but the rhino skin is more like finding faces in clouds. Yeah, some of the cracks seem similar, but they aren’t near each other and what patterns that do appear, seem erratic. In contrast, the reptile scales have an almost “factory stamped” look to them. Another noticeable difference is in the sharp borders seen between each scale. Compare that to the very diffuse borders seen in the skin of the rhinos above.

Paleontologist, In Sung Paik, provided a set of criteria to determine scales from other biological / geological structures (Paik et al. 2010). These criteria were:

  1. Pattern — Dinosaur scales show interlocking polygons in a regular pattern
  2. Contact  & interlocking state— Edge-to-edge contact typically found in scales (more distant contact possible if patterned)
  3. Edge shape — Dinosaur scales show pointed edges (i.e., each scale is distinguishable from the next)
  4. Internal microfabric of the scales — Scales show an internal microtexture to them (dependent on degree of preservation)

We can see that the three reptile scales above fit these criteria, whereas the rhino skin imprints do not.

Similarly, Wyrex’s scale impressions also fullfill Paik et al.’s criteria. Image adapted from Fig. 1B in Bell et al. 2017.

Another clear sign that the impressions found for these tyrannosaurids, are likely scales, comes from the detailed description by Bell et al. on the integument impressions of Albertosaurus sarcophagus.

Most notable, however, is a specimen of Albertosaurus sarcophagus (TMP 1994.186.0001) that preserves evidence of feature scales on the abdomen. The feature scales are conical (7 mm in diameter, 2.5 mm high) with radial corrugations and are embedded in a patch of pebbly, circular to hexagonal basement scales (diameter range = 1.4–2.5 mm; electronic supplementary material, figure S1a–d).

You don’t get feature scales on naked skin. It just doesn’t happen.

Iguana neck scales. See those large tubercular scales surrounded by smaller ones? Those are feature scales. Image by John Daniels

So yeah, the whole “naked, wrinkly skin” argument doesn’t really work for these integument impressions.

Okay, so what about the other go-to argument? How much do tyrannosaurid skin impressions resemble a plucked chicken?

I have to admit, I don’t get it. If we use the criteria from Paik et al. 2010, we can see that a plucked chicken does satisfy criteria one, in that the plucked areas do show a distinct pattern. However, everything falls apart after that, as feathered areas do not show any type of interlocking polygonal pattern. Instead, the plucked areas look more like raised ridges of skin, which they essentially are. Further, chicken feathers, like those of all birds [most birds], grow in distinct tracts. This is quite prominent in the image to the right, where one can see the feathers taking up distinct sections of the skin, with other areas being completely devoid of feathers. If we were seeing this scenario in tyrannosaurids, we might expect to see large smooth areas of skin between feather tracts. This would be all the more likely if the filaments seen in early coelurosaurs represented protofeathers.

I’ve often heard it said that many dinosaur skin impressions look like plucked chickens, but do they really? Image is unattributed, but widely available online.

Of course, it’s also possible that the limitation of feathers to distinct regions of the body happened later in avian evolution [note: ratites do not show feather tracts, suggesting that this may be more of a neognath synapomorphy]. Evo-devo work does indicate that featherless regions of the body can be induced to form feathers without much effort (Dhouailly 2009), and a recently accepted manuscript describing a young enantiornithine trapped in amber, does suggest that early bird lineages did have more feathers than extant birds (Xing et al. 2017, accepted manuscript), so the possibility remains that protofeathered dinosaurs did not limit their protofeathers to distinct tracts.

Nonetheless, the fact remains that we still do not see any interlocking pattern with pointed edges, nor any scale-like microtexture. In fact, the texture in bird skin is pretty much completely smooth.

There is another problem with the “plucked chicken” argument. It requires that the animals get denuded prior to burial. The chicken shown in the image to the right is not normal. A filament-covered theropod that dies in a fossil-preserving situation is unlikely to have the exposure time necessary for filaments to slough off, and if it does, we would expect to see some type of folds or wrinkles associated with the sloughing. Another thing to consider is the remarkably thin skin of birds (Pass 1995), which would make plucked skin even less likely to preserve, much less show any type of texture. Lastly, even if an animal like Wyrex or Daspletosaurus did have its integument preserved in this weird fashion, it would be an incredibly rare find, and not one that would likely be repeated in other tyrannosaurid integument impressions.

So the plucked chicken idea doesn’t really hold water.

Laslty, we have dirt, or other simple sedimentological structures.

Fossilized mudcracks in positive (left), and in negative (right). Left image by C.E. Jones. Right image by Greg Wiles

At first glance, these structures certainly have the most potential to be mistaken for scales. They have a tight interlocking polygonal pattern just like dinosaur scales, but the individual polygons are still randomly arranged. They also lack any fine microtexture to them, since they are just mud. Nontheless, the potential to be confused with scales is still pretty strong here. Paik et al. 2010 even mentioned this potential confusion and cautioned that some published accounts of scales may in fact be other biological or geological structures. Thankfully, the biggest worry of confusing these structures with scales happens when the scales are found in isolation. None of the specimens described by Bell et al. 2017, were found in isolation, save for the Daspletosaurus scale impressions, and even that is a maybe. Given that the specimen that these impressions came from (TMP 2001.036.001) is currently missing, it is hard to determine whether these scales were found in isolation or not. Still, that is only one of five different specimens, and even then the scalation appears to fit the description of the other tyrannosaurids.

So the argument for sedimentological structures, while important to consider, can likely be ruled out due to the proximity of the scales to underlying bones, as well as the feature scales found in A. sarcophagus.

(3) The authors are making blanket statements based off coin-sized samples of “skin”

Composite of all known tyrannosaurid skin impressions, their size, and their location on the body. Collaborative image by Joshua Balize

While I am the first person to jump on a paper for making broad statements that extend beyond the data, I do not think that this accusation is warranted in this case. The authors set out to first describe the shape and distribution of scales in tyrannosaurids. The most I can fault them for is their use of Tyrannosauroid in their paper title. However, I suspect this had more to do with incorporating Dilong paradoxus and Y. huali more than anything else.

I have also heard complaints that the authors didn’t bother to describe the integument found in D. parodoxus and Y. huali, but to that I would say: read the original papers. Both Xu et al. 2004 and 2012, went into as much detail as possible in their descriptions of the filaments found in these two taxa. A further analysis from the authors would not have added much to the conversation.

In contrast, scales have always been given the short end of the stick when it comes to popularization and description. Phil Bell and colleagues represent a new guard of paleontologists that realize that scales were an equally important, if not more important aspect of many dinosaur lives. The work by Bell et al. 2017 stems from previous work on hadrosaurs that included the first ever classification of scale terminology in dinosaurs (Bell 2012), including their use as diagnostic tools for species identification. This attention to detail on scale morphology made these authors the ideal paleontologists to tackle tyrannosaurid integument in a way that removes much of the worry about accidentally assigning scales to non-biological structures.

As for the “coin-size” argument, for that I would like to turn your attention to the incredible work done by Joshua Ballze and company. While it’s true that the combined total for these scale impressions is still pretty small by tyrannosaurid standards, they are a far cry from being just “coin sized”, or even laptop sized pieces of integument.

Further, in spite of their small overall size, these skin impressions represent almost every major part of the body. Thus, when Bell et al. argued that tyrannosaurids were likely scaly all over, it is based on the reasonable assumption that the scales on the tail met up with the scales on the legs, belly, abdomen, neck, and head. This is not extending one’s interpretations beyond the data, but rather using reasonable assumptions and applying them to known integument distributions in these taxa.

Another important aspect to consider is that the authors took their work a step further and performed an ancestral character reconstruction based on known integument distributions in tyrannosauroids. They found that scaly tyrannosaurids (i.e., devoid of any filaments) have a 97% probability of being true, despite an 89% probability that tyrannosauroids started off filamented. The authors arguments for why tyrannosaurids would revert back to scales are interesting and worth pursuing. For instance, if the reversion was driven by body size, then why did Y. huali retain its shag, whereas similarly sized A. sarcophagus was scaly? If climate was responsible for shaggy retention in Y. huali, then why did A. sarcophagus, and Daspletosaurus not retain filaments in their equally “cool” environments?

More importantly, why should scales re-evolve at all? Why not just bare skin? Hopefully this study will encourage others to look into why scales evolved in the first place, and actually test some of their functions rather than stick with the party line of “water retention” and “abrasion protection”.

Some notes on scale size

It’s worth discussing the size of the scales found in tyrannosaurids. Namely, that the scales are remarkably small. The average scale size from all the specimens studied by Bell et al., is ~2.2 mm (range: 1–4.9 mm, not counting feature scales). For reference, this link provides examples of millimeter-sized everyday structures. So we are talking about some remarkably tiny scales. The small size of these scales are what lead paleontologists like Phil Currie, to state that tyrannosaurids were truly naked (i.e., scaleless). So, although tyrannosaurid integument impressions show scales, the scalation in tyrannosaurids would have been remarkably fine.

All that said, a comparison with other dinosaur groups suggests that fine scalation is actually pretty common place within Dinosauria.

The stegosaur, Hesperosaurus mjosi, preserves scales that range from 2–7 mm in diameter (Christiansen et al. 2010), whereas Gigantspinosaurus sichuanensispreserved slightly larger scales at 5.7–9.2 mm (Xing et al. 2008). Ankylosaurids are known for their incredibly well-developed osteoderms, but well-preserved specimens also show that these osteoderms were surrounded by much smaller basement scales that ranged from 1.3–5 mm in animals like Pinacosaurus grangeri (Gilmore 1933) and an undetermined ankylosaur (Arbour et al. 2014). Hadrosaurs are certainly the most common dinosaurs to have presrved integument (Davis 2012), yet even in this group the basement scale range is still small at 1–10 mm in diameter (Bell 2012). Even sauropods, despite their immense sizes, are known to preserve remarkably small scales, such as the small (1–3 mm diameter) type II (= basement) scales in Tehuelchesaurus benitezii (Gimenéz 2007), as well as another indeterminate sauropod with 1–2 mm width basement (Foster & Hunt-Foster 2011). It appears that basement scales in dinosaurs were pretty conservative in size between taxa. Feature scales vary much more and could reach up to 102 mm (Sternberg 1925), before accounting for osteoderm scutes. Yet even here, the few preserved feature scales in A. sarcophagus, fall within the minimum size range for hadrosaurs like Saurolophus osborni (Bell 2012).

Compared to other dinosaurs, tyrannosaurids fall on the low end of average. I suspect that the rather small individual samples that we have may also be biasing our interpretations in tyrannosaurids. As at least one A. sarcophagus specimen shows, there was regional variation along the body. Unfortunately, without a true tyrannosaurid mummy, we won’t know just how varied these scale patterns were in this group.

Ventral scales on a 2 meter Crocodylus moreletii. Scales are an “enormous” 45 mm. Image from Fig 1 of Tellez and Paquet-Durand 2011

I suspect that crocodile scutes and ceratopsian skin impressions may be skewing public perception of what dinosaur scales should look like. In crocs, very large scales predominate the dorsal and ventral aspects of the body, whereas ceratopsian scale impressions tend to be dominated by large (5 cm or more) scales (e.g., Sternberg 1925), though even in Ceratopsia, basement scales still tend to remain pretty small at around 3.2 mm in diameter (Davis 2012, supplemental table).

These large scales, along with the fact that many dinosaurs were very enormous animals, tends to skew our perceptions towards large-scalation in large dinosaurs. Looking at the data, however, suggests that large scalation was more limited to specific taxa and specific scale types (i.e., feature scales). Many dinosaurs appear to have grown large while maintaining fairly small scales.

This pattern is more consistent with what we see in lepidosaurs than crocodylians. Compare, for instance, the scale size in land iguanas (Conolophus, Gentile and Snell 2009) or reticulated pythons (Python reticulatus, Auliya et al. 2002). Both are large animals in their respective lineages, yet their scale sizes (aside from head scales) remain remarkably small.

Illustrations of head and neck scalation in three subspecies of P. reticulatus ranging from 1.3–1.7 meters in total length. Image from Fig 3 of Auliya et al. 2002.

I think we tend to see lizard scales as larger than they are because many detailed lizard pictures are taken up close. Case in point is the study from earlier this year on a new species of fish-scaled geckos (Geckolepis). The pictures in the original paper (Scherz et al. 2017) and associated news articles, makes the animals appear much larger than they are. In the images, the new species, Geckolepis megalepis, is shown wearing a seemingly oversized coat of scales. Yet, reading the paper we find that the largest scales found on the species were 5.8 mm long, which is smaller than the feature scales on A. sarcophagus. What’s nearly indescribably tiny on a tyrannosaurid, is almost 10% of the snout vent length on this little gecko.

The detachable, nearly 6 mm scales on Geckolepis megalepis, seem almost overgrown on the animal. Image from Fig 3A of Scherz et al. 2017

It would be interesting to see how scalation changes with body size, but to date no one has studied this. It’s likely contingent on phylogeny (you use what you have), but I wouldn’t be surprised to find that scale count, rather than scale size, increases with body size in lepidosaurs and dinosaurs.

Enfluffening the gaps

Lastly, I want to talk about the other thing I have noticed in relations to these papers. I saw this pop up a little bit with the Carr et al. paper, and a whole lot with the Bell et al. paper. Opponents to the idea of scaly tyrannosaurids argue that , even if the scales found on the body / face are true:”They could still have feathers on the back / between the scales / when they were young“.

This argument reminds me a lot of the God of the Gaps argument used in creationism vs evolution debates. This argument proposes that anywhere where science lacks sufficient knowledge of something, that is where god lies. The problem with this approach is that as science explains more and more of the natural world, the influence of god on nature gets pushed more and more to the periphery. Thus, god just becomes a placeholder for areas that we currently don’t know. We appear to be seeing similar arguments being put forth for the presence of filaments in dinosaurs.

I’ve heard it argued for sauropods, despite embryonic remains showing scales. I even seen it used on ornithopods, despite near complete scaly mummies. Now, with tyrannosaurids, this “enfluffening the gaps” is being proposed for the back of the animals (the only area that we lack integument preservation) and for hatchlings.

Points to the authors, Bell et al. do tackle these enfluffening arguments in the original paper.

Combined with evidence from other tyrannosaurids…provides compelling evidence of an entirely squamous covering…suggesting that most (if not all) large-bodied tyrannosaurids were scaly and, if partly feathered, these were limited to the dorsum.

Finally, the presence of epidermal scales in a large adult individual does not rule out the possibility that younger individuals possessed feathers—a developmental switchover that, to our knowledge, would be unprecedented at any rate.

This line of reasoning essentially echoes what I have written before. Ontogenetic shifts in integumentary appendage type does not happen in any extant animal. As such, proposing that it happened in dinosaurs constitutes an extraordinary claim that would require extraordinary evidence. As for that infamous feathered mohawk trope, I have also expressed my problems with that argument before.


These two new papers on tyrannosaurid integument were a long time coming, and regardless of one’s take on the data, it is great to see these specimens get a real description. That both papers have received such heated interactions will hopefully spur other paleontologists to look for more soft tissue specimens, and even do some ultrastructural analyses of preserved scales to see how much of the origiinal morphology survived. Who knows, maybe one day we will finally find that elusive tyrannosaur mummy.

Regardless, the data presented in these two papers certainly indicates that scales are important integumentary structures that warrant the same level of scrutiny and analysis that filaments have been getting since 1996.

~ Jura


Arbour, V.M., Burns, M.E., Bell, P.R., Currie, P.J. 2014. Epidermal and Dermal Integumentary Structures of Ankylosaurian Dinosaurs. J. Morphol. 275(1):39–50.
Auliya, M., Mausfeld, P., Schmitz, A., Böhme, W. 2002. Review of the Reticulated Python (Python reticulatus Schneider, 1801) with the Description fo New Subspecies from Indonesia. Naturwissenschaften. 89:201–213.
Barker, C.T., Naish, D., Newham, E., Katsamensi, O.L., Dyke, G. 2017. Complex Neuroanatomy in the Rostrum of the Isle of Wight Theropod Neovenator salerii. Scientific Reports. 7:3749.
Barrett, P.M., Evans, D.C., Campione, N.E. 2015. Evolution of Dinosaur Epidermal Structures. Biol. Lett. 11:20150229.
Bell, P.R. 2012. Standardized Terminology and Potential Taxonomic Utility for Hadrosaurid Skin Impressions: A Case Study for Saurolophus from Canada and Mongolia. PLoS One. 7(2):e31295.
Bell, P.R., Campione, N.E., Persons, W.S., Currie, P.J., Larson, P.L., Tanke, D.H., Bakker, R.T. 2017. Tyrannosauroid Integument Reveals Conflicting Patterns of Gigantism and Feather Evolution. Biology Letters. 13:20170092.
Brusatte S.L., Carr T.D. 2016 The Phylogeny and Evolutionary History of Tyrannosauroid Dinosaurs. Scientific Reports 6:20252.
Carr, T.D., Varricchio, D.J., Sedlmayr, J.C., Roberts, E.M., Moore, J.R. 2017. New Tyrannosaur with Evidence for Anagenesis and Crocodile-Like Facial Sensory System. Scientific Reports. 7(44942):1–11.
Christiansen, N.A., Tschopp, T. 2010. Exceptional Stegosaur Integument Impressions from the Upper Jurassic Morrison Formation of Wyoming. Swiss. J. Geosci. 103(2):163–171.
Currie, P.J., Badamgarav, D., Koppelhus, E.B. 2003. The First Late Cretaceous Footprints from the Nemegt Locality in the Gobi of Mongolia. Ichnos. Vol.10:1-12.Currie, P.J., Badamgarav, D., Koppelhus, E.B. 2003. The First Late Cretaceous Footprints from the Nemegt Locality in the Gobi of Mongolia. Ichnos. Vol.10:1-12.
Currie, P.J., Koppelhus, E.B. 2015. The Significance of the Theropod Collections of the Royal Tyrrell Museum of Palaeontology to our Understanding of Late Creataceous Theropod Diversity. Can. J. Earth. Sci. 52:620–629.
Davis, M. 2012. Census of Dinosaur Skin Reveals Lithology May not be the Most Important Factor in Increased Preservation of Hadorsaurid Skin. Acta. Paleo. Polo. 59(3):601–605.
Foster, J.R., Hunt-Foster, R.K. 2011. New Occurrences of Dinosaur Skin of Two Types (Sauropoda? and Dinosauria indet.) from the Late Jurassic of North America (Mygatt-Moore Quarry, Morrison Formation). J. Vert. Paleo. 31(3):717–721.
Gentile, G., Snell, H. 2009. Conolophus marthae sp. nov. (Squamata, Iguanidae), A New Species of Land Iguana from the Galápagos Archipelago. Zootaxa. 2201:1–10.
George, I.D., Holliday, C.M. 2013. Trigeminal Nerve Morphology in Alligator mississippiensis and its Significance for Crocodyliform Facial Sensation and Evolution. Anat. Rec. 296(4):670–680.
Gilmore, C.W. 1933. Two New Dinosaurian Reptiles from Mongolia with Notes on Some Fragmentary Specimens. Am. Mus. Nov. 679:1–20.
Gimenéz, O.V. 2007. Skin Impressions of Tehuelchesaurus (Sauropoda) from the Upper Jurassic of Patagoinia. Rev. Mus. Argentino Cienc. Nat. n.s. 9(2):119–124.
Hieronymus, T.L., Witmer, L.M., Tanke, D.H., Currie, P.J. 2009. The Facial Integument of Centrosaurine Ceratopsids: Morphological and Histological Correlates of Novel Skin Structures. Anat. Rec. 292:1370–1396.
Ibrahim, N., Sereno, P.C., Dal Sasso, C., Maganuco, S., FAbbri, M., Martill, D.M., Zouhri, S., Myhrvold, N., Lurino, D.A. 2014. Semiaquatic Adapations in a Giant Predatory Dinosaur. Science Express. 345(6204):1613–1616.
Milinkovitch, M.C., Manukyan, L., Debry, A., Di-Poi, N., Martin, S., Sing, D., Lambert, D., Zwicker, M. 2013. Crocodile Head Scales are not Developmental Units but Emerge from Physical Cracking. Science. 339:78–81.
Paik, I.S., Kim, H.J., Huh, M. 2010. Impressions of Dinosaur Skin from the Cretaceous Haman Formation in Korea. J. Asian Earth Sci. 39:270–274.
Pass, D.A. 1995. Normal Anatomy of the Avian Skin and Feathers. Seminars in Avian and Exotic Pet Medicine.4(4):152–160.
Scherz, M.D., Daza, J.D., Köhler, J., Vences, M., Glaw, F. 2017. Off the Scale: A New Species of Fish-Scale Gecko (Squamata: Gekkonidae: Geckolepis) with Exceptionally Large Scales. PeerJ. 5:e2955.
Sternberg, C.M. 1925. Integument of Chasmosaurus belli. Canadian Field Naturalist. 39:108–110.
Tellez, M., Paquet-Durand, I. 2011. Nematode Infection fo the Ventral Scales of the American Crocodile (Crocodylus acutus) and Morelet’s Crocodile (Crocodylus moreletti) in Southern Belize. Comp. Parasitol. 78(2):378–381.
Witmer, L.M., Ridgely, R.C. 2009. New Insights into the Brain, Braincase, and Ear Region of Tyrannosaurs (Dinosauria, Theropoda), with Implications for Sensory Organization and Behavior. Anat. Rec. 292:1266–1296.
Xing, L., O’Connor, J.K., McKellar, R.C., Chiappe, L.M., Tseng, K., Bai, M. 2017. A Mid-Cretaceous Enantiornithine (Aves) Hatchling Preserved in Burmese Amber with Unusual Plumage. Gondwana Research. Accepted manuscript.
Xing, L., Peng, G., Shu, C. 2008. Stegosaurian Skin Impressions from the Upper Jurassic Shangshaximiao Formation, Zigong, Sichuan, China: A New Observations. Geol. Bull. China. 27(7):1049–1053.
Xu, X., Norell, M.A., Kuang, X., Wang, X., Zhao, Q., Jia, C. 2004. Basal Tyrannosauroids from China and Evidence for Protofeathers in Tyrannosauroids. Nature. 431:680–684.
Xu, X., Wang, K., Zhang, K., Ma, Q., Xin, L., Sullivan, C., Hu, D., Cheng, S., Wang, S. 2012. A Gigantic Feathered Dinosaur form the Lower Cretaceous of China. Nature. 484:92–95.


Facebooktwitterredditpinterestlinkedintumblrmailby feather

28 Responses to The return of the scaly T. rex to modern paleo-art

  1. Pingback:Eric Olson

  2. Pingback:Brandon Pilcher

  3. Pingback:Nick

  4. Pingback:Duane Nash

  5. Pingback:Mat

    • Pingback:Jura

      • Pingback:Mat

  6. Pingback:Jura

  7. Pingback:Brandon Pilcher

  8. Pingback:MrCrow

  9. Pingback:MrCrow

  10. Pingback:Herman Diaz

  11. Pingback:MrCrow

  12. Pingback:Deltadromeus

  13. Pingback:MrCrow

  14. Pingback:MrCrow

  15. Pingback:Crazy Z

  16. Pingback:MrCrow

  17. Pingback:Joe Daniel

  18. Pingback:Thiago Baz