Sunday, September 1, 2013

You're Doing It Wrong : Dino Foot Scales

Above: Our subject matter.
It's often said by those who support a strict phylogenetics-based system of naming life that it's only by restricting well-known names from neontology (the study of modern organisms) to crown groups can we avoid making unjustified assumptions about members of stem-groups.

These kinds of unjustified assumptions have been rampant in the history of studying stem-birds. Archaeopteryx has traditionally been depicted, incorrectly, with a reversed hallux, and occasionally even with beak-like structures, simply because it's a "bird", and those are features all birds have. Except Archaeopteryx is not a true "bird", it's a stem-bird, more closely related to birds than to any other living animal group, but not a member of the group that includes all modern birds. It's fair to assume that an extinct member of the duck lineage, like Vegavis, had a bill, but that's not necessarily so for, say, Patagopteryx, despite the fact that it is usually referred to as a "bird".

Modern bird feet, by Philip Henry Gosse, 1849, public domain. Note overlapping scutes on
the top surfaces, and pebbly, polygonal reticulae on the bottom surfaces.

Most paleoartists have absorbed these kinds of warnings, and do a good job of avoiding obvious errors based on typology, the assumption that all species in a certain "type" share "key characteristics." But there are some typological memes in the bird lineage that are more pernicious, possibly because their actual evolution is something most artists don't think about very much.

Take, for example, the bird-like scutes that are almost universally illustrated covering the tarsus (upper foot/lower hind limb) of dinosaurs. Is there any evidence that these were actually present in any given group of non-theropod stem birds? Well... no. Not that I'm aware of (if you know differently, please comment!).

Sinosauropteryx prima with tarsal scutes.
Image by Matt Martyniuk,  licensed.
I'm not sure when this meme began, and if it's related to the Dinosaur Renaissance when the link between birds and dinosaurs was re-established. Looking at some Charles Knight paintings, such as his famous "Leaping Lealaps", it appears that the feet of his theropods were scaled based on modern lizards (more on the differences between lizard scales and other types of "scales" below). Bakker's influential early restoration of Deinonychus does not include any obvious scutes on the feet or tarsus. Mark Hallet, on the other hand, did include what look like oblong bird-like scutes on his theropods. At any rate, it's hard to deny that "bird feet" are typical of almost all modern reconstructions of dinosaurs, including my own, and are not limited to theropods. Bird-feet are often restored on ornithischians and even pterosaurs.

Of course, like many paleo-memes that developed during the 1980s, the main idea seems to be using this as a flourish to make otherwise scaly dinosaurs seem more bird-like. And thanks to skin impressions, we know that many dinosaurs had scales, right?

The problem here is that people tend to lump any scale or scute-like dermal structure that's not a hair or a feather into the category of "scale", despite long-documented and rather drastic differences in their structure and development.

Tarsal scutes
Tarsal scutes of a seriama. Photo by Matt Edmonds, licensed.
Let's survey the body covering of modern birds. Birds actually have three basic types of integument: feathers (of course, though they come in a variety of forms which themselves develop differently), reticula, and scuta.

Scuta are the broad, flat, semi-overlapping "scales" on the top of the tarsus and the toes. The development of these structures is controversial, but as many of you may know, they might in fact be derived from feathers themselves, and not related to the true scales found in lizards and snakes at all (Dhouailly, 2009). Rather famously, the scuta of bird feet can easily be converted partially to feathers in experiments (the tips become feathery, while the anterior portions stay scute-like) and in natural mutations, as in some chickens.

Skin impressions from the foot of Saurolophus osborni, showing
only reticulae. From Bell 2012, licensed (CC BY).

More interestingly, in mutations where the scuta become feathers, they specifically become remiges (wing feathers) or remix-like feathers. It appears not to be possible to induce the reverse (feathers on the body to become scuta). Researchers like Dhouailly (2009) have suggested that these scuta might be homologous with the hind-wings feathers of basal paravians like Microraptor, and that the ancestral condition prior to the evolution of scutes from foot remiges as a fully- or nearly-fully feathered foot. The only non-feather skin impressions from non-avialan maniraptorans that I know of come from the "Dave" specimen of Sinornithosaurus. This preserves reticula (see below) on the plantar surface of the digits, but not scuta. It's very possible the rest of the foot was covered in feathers, with reticular "pads" for walking (think the foot pads of dogs). Skin impressions from hadrosaurids like Saurolophus demonstrate that bird-like tarsal scutes were absent, and that the entire surface of the metatarsus and digits was covered in reticulae (see photo above).

(The bird-foot meme is so ingrained that I found some blogs reporting that these specimens actually had tarsal scutes, even speculating on their purpose, despite the paper clearly contradicting the idea!)

Reticula, or "Tuberculate scales"
Feet of a Little Egret. Photo by J.M. Garg, licensed.

Reticula, often referred to as "tuberculate scales", are the bumpy structures on the plantar side of the foot. These are also, famously, the type of "scales" found in just about all non-feathered stem-bird skin impressions. And here's the kicker--these have very little to do with the "scales" of squamates or even crocodiles.

In modern birds, the reticula, unlike feathers or scutes, do not form from placodes. Unlike lizard scales, scuta, and feathers, they're made not of beta-keratin, but alpha-keratin (the same as mammalian hair and amphibian 'warts'). The reticulae do start with an overlying beta-keratin layer that is shed during development. While they don't form from placodes, they appear to be similar to the pre-placode stage of feather formation. Reticula apparently never morph into scales in nature (I suppose ptarmigans start with naked skin on their feet?), this can be induced in the lab. Dhouailly suggests that, given the weight of the developmental evidence, reticula appear to be feathers which are arrested in a very early stage of development.

You read that right--tuberculate "scales" are probably modified feathers! So, those fossilized hadrosaurid mummies covered in "scales"? It's possible or even likely that they had feathered ancestors, and evolved in such a way to stop the development of feathers early on, instead yielding a fine, pebbly skin covering derived from old feather genes.

Dhouailly, in her phylogeny of amniote skin coverings, supposes that both feathers and crocodilian scutes evolved from a hypothetical, primitive, amphibian-like granulated skin covering consisting of alpha-leratin, which also have rise to mammalian hair. Tuberculate scales and bird scuta are suggested to both be derived from feathers secondarily, with reticula only being superficially similar to the ancestral, 'warty' condition. Dhouailly shows non-theropod dinosaurs in her diagram to have croc-like scutes, but most reported skin impressions from dinosaurs actually show reticula instead. The exception may be the large osteoderm scutes of thyreophorans and the reported larger scute-like scales and croc-like belly scales of Triceratops, though given the variety in bird "scales", these could easily be feather derivatives as well. In all cases of dinosaurs with osteoderms, they seem to be interspersed among tuberculate scales the same way the scutes of bird feet are, rather than being arranged in neat rows bordered by bare skin as in crocs.

Dino Feet
Feet of a Great Horned Owl. Note the dorsal toes and tarsals are mostly feathered
rather than covered in scutes, with the ventral side covered in reticulae, probably
analogous to the ancestral condition for at least coelurosaurs but possibly all ornithodirans.
Photo by Andrea Westmoreland, licensed.
So, back to the main premise here: what does this mean for artists? The fact that tarsal scutes appear to be derived from feathers, and from pennaceous feathers in particular, suggest they are probably restricted to at most Tetanurae, given the current phylogenetic bracketing. However, numerous maniraptoran fossils preserve fully-feathered hind limbs, including toes but probably excluding foot pads, which were covered in reticulae (e.g. Sinornithosaurus). The most basal stem-birds with evidence of tarsal scutes are Yanornis (Zheng et al. 2013) and Concavenator, with some fluffy-footed forms in between, suggesting that once developed, the genetic 'switch' between feathered and scutted feet may have been highly variable, as it is in modern birds.

In feathered species, the feathers on the bottom of the foot seem to have been developmentally changed to reticula for better traction/less damage to feathers. In some large species, fossorial species, semi-aquatic species, etc., the reticula probably spread to other parts of the body. The general feather-derived skin covering also could have produced the various other kinds of skin covering, like psittacosaur "quills" (Dhouailly, 2009).

In the end, while the paleo-art meme of scuty tarsals in non-theropod stem-birds was invented to make the link between dinosaurs and birds more obvious, it turns out that it may actually be too bird-like for many of them. Hopefully, another unjustified assumption about the stem-group made by copying a trait from the crown can go the way of the ornithischian.

As for the evolution of scales and feathers and the different kinds and patterns of skin coverings in stem-birds, at the very least it should be clear that a simple dichotomy of scales vs. feathers is about as far from reality as you can get. Feathers could become "scales", scale-like scutes could become feathers, and novel structures of all kinds can appear, derived from any of them (like the beards of turkeys, which are derived from alpha-keratin and are not true feathers, but some kind of derivative like tuberculate scales [Sawyer & Knapp, 2003]). Each different lineage of stem-birds may have been doing very different things with the very adaptable basic feather or feather-like structures made possible by their crazy dermal genes.

UPDATE: As I figured I might, I overlooked a major data point here--Concavenator (thanks to Tom Parker in the comments for pointing this out!). Concavenator is an interesting case that shows these type of things are not always cut and dry, since it does preserve scutes on the tarsals and toes, and scutes on some other parts of the body including the underside of the tail.* So, if Dhouailly 2009, Zheng et al. 2013, etc. are correct about the development and evolution of tarsal scutes from feathered feet, where does this leave Concavenator? A few possibilities: one, the developmental data is wrong, and tarsal scutes are precursors to pennaceous feathers rather than the other way around. Two, the tarsal scutes of Concavenator are not homologous with those of birds. Three, Concavenator evolved from ancestors with feathers or feathered feet. I find the third option to be the most likely right now. Notoriously, Concavenator was described with structures resembling quill knobs on the arm, which in eumaniraptorans anchor the sort of pennaceous feathers developmental data suggests are the precursors of tarsal scutes. While the identification of these knobs has been challenged, the fact that tarsal scutes exist in this species may be further support for the idea that Concavenator did indeed have remixes (or that its ancestors did).

* I originally wrote "reticulate scales on the rest of the body", but this is not correct--scutes, not reticulae, are preserved in impressions other than the pes.

Coming real soon, much more on a particularly cool branch of stem-birds, in the form of the amazing new book by Mark Witton...


Bell, P. R. (2012). Standardized terminology and potential taxonomic utility for hadrosaurid skin impressions: a case study for Saurolophus from Canada and MongoliaPloS one7(2), e31295.

Dhouailly, D. (2009). A new scenario for the evolutionary origin of hair, feather, and avian scalesJournal of anatomy214(4), 587-606.

El-Sayyad, H. I., Fouda, Y. A., Khalifa, S. A., AL-Gebaly, A. S., & El-Sayyad, O. K. (2013). Studies on epidermal appendages of chick embryosInt. J. Curr. Microbiol. App. Sci2(5), 315-327.

Sawyer, R. H., & Knapp, L. W. (2003). Avian skin development and the evolutionary origin of feathers. Journal of Experimental Zoology Part B: Molecular and Developmental Evolution298(1), 57-72.

Zheng, X., Zhou, Z., Wang, X., Zhang, F., Zhang, X., Wang, Y., ... & Xu, X. (2013). Hind wings in basal birds and the evolution of leg feathersScience,339(6125), 1309-1312.


  1. "Take, for example, the bird-like scutes that are almost universally illustrated covering the tarsus (upper foot/lower hind limb) of dinosaurs. Is there any evidence that these were actually present in any given group of non-avialan stem birds? Well... no. Not that I'm aware of (if you know differently, please comment!)."

    I do know differently. They are present on _Concavenator_.

    "The scale diversity shown by Concavenator is quite similar to that of some present-day Neornithes (modern birds), in which reticulate, scutate and scutella scales are present. The former correspond to the footpads. Scutate scales are rectangular and are situated on the anterior metatarsal area and the dorsal digit zone. Scutella scales are placed lateral to the scutate scales; they are smaller but also rectangular in outline."

    A bizarre, humped Carcharodontosauria (Theropoda) from the Lower Cretaceous of Spain (Ortega et al. 2010)

    1. Gag, knew I'd overlook one! Thanks, I'll update the post when I get a chance.

    2. Tom preceded me.
      Tubercolate scales are also known in Hesperornithes feet (which is usually placed closer to crown-birds than Yanornis, although my most recent analysis with the inclusion of Piscivoravis found its relative Hesperornis outside the Aves-Songlingornithidae node).
      Ref: Williston, S.W. 1896.On the dermal covering of Hesperornis. Kansas University Quarterly 5(1):53-54, pl. II.

    3. "my most recent analysis with the inclusion of Piscivoravis found its relative Hesperornis outside the Aves-Songlingornithidae node"

      "Tubercolate scales are also known in Hesperornithes feet "
      True, and scuta as well on parts of the distal tarsus, though other parts of the tarsus were feathered at least in Parahesperornis.

  2. Birds vary in the expression of scutes, tubercules, and scutellae on their feet, some bearing only the second, others interspersing the second with the first, and varying in how much of the third are also present or, how many have true feathers on their feet. That crocs have large scales and so do birds, if they have the same development and structure then there is no doubt that some (if not most or all) extinct nonavian, even feathered, dinosaurs had them, too. But they could vary amongst these taxa just the same.

    1. It seems, though, that croc scutes and bird scutes don't share the same development, though they're derived from the same precursors, with bird scutes being derived from feathers and croc scutes and feathers both being derived from

  3. "(The bird-foot meme is so ingrained that I found some blogs reporting that these specimens actually had tarsal scutes, even speculating on their purpose, despite the paper clearly contradicting the idea!)"

    My bad, I guess I misunderstood the paper during my first time reading through it. I noted the mistake on the post. Thanks for pointing it out. ;)

  4. Never knew that about Saurolophus' foot scales.

    While I have no issue with dinosaur non-overlapping epidermal derivatives being reticula, and thus being derived from feathers, what's the evidence they're not just tuberculate scales as in crocodilians?

    1. Good question, and it might not be one we can test without genetics. WE need a better comparison of these various structures rather than just lumping them all together as "scales" and calling it a day, as is usually done in descriptions of skin impressions.

      That said, do any crocodilians have tuberculate scales of the kind seen in dinosaur skin? Most close-ups of croc skin I can find show even the small pavement scales in crocs clearly scute-like (similar to scutella in birds), not the kind of pebbly texture or radial patterns typical of stem-birds and birds. Dino skin at least "looks" a lot closer to bird reticulae than croc scutella, superficially (which may be the best we can do).

    2. I think croc skin might be equally heavily derived though. What we need is skin impressions from some basal archosaurs, or even some early Pseudosuchians or Dinosauromorphs.

      This is clearly only wishful thinking at this stage, but if we knew this we could see more clearly if Neornthischians, sauropods, _Carnotaurus_, etc. are showing the basal condition or all convergently developing similar derivatives.

  5. Excellent post! I have always had difficulty in being able to choose the type of integument of the feet in my sketches, but now I can be safer.

  6. Wonderful post, really usefull.
    After hearing from various sources that a feather cannot become a scale (without citing any study or proof), this certainly is something 'new'. I always thought that feathers were a common feature of all dinosaurs and now this idea is even stronger than ever. Again, thanks for sharing.

    Hope to read your pterosaur post soon.

    1. Thanks! As I mentioned in the post, it's true that modern feathers cannot be induced to become scutes. But the genetic evidence suggests that this must have happened at some point in bird ancestry.

  7. One more small point about _Concavenator_, post-update:
    "it does preserve scutes on the tarsals and toes, and reticular on most of the body"

    As far as I can see in the paper, there is no mention of reticulate scales on the body. I may be unaware of some additional information (perhaps Andrea knows) but the paper only seems to report integument on the digits (incl. foot pads), metatarsal and the underside of the tail (in the form of rows rectangular scutes like _Triceratops_).

    As far as I know via. Ortega et al. there is no other integument preserved.

  8. "Archaeopteryx has traditionally been depicted, incorrectly, with a reversed hallux,"

    To be fair, the hallux IS reversed ("Only weakly reversed", yes, but reversed nonetheless: ).

  9. Pterosaur also have reticula like scales on their foot pads, with no scutes known. Maybe these "scales" are derived pycnofibrils? At any rate, it shows that scutes probably were not ancestral to the last common ancestor between pterosaurs and birds.

  10. With regards to the origin of the Paleo-art meme, I clearly remember playing with a pre-Renaissance T. rex toy that had tarsal scutes on its feet; they were distinct enough that I even remember my 8-year-old (or whatever) self making the comparison to a bird. I sadly don't have the toy now, or I'd include a picture. This certainly isn't to suggest that it represents the origin or even a reference to the origin of the meme, as it may have been an independent decision on the part of the manufacturer, but I figured it was notable nonetheless.

  11. There seems to be a lot of extant lizards and reptiles with some form of overlapping scales on their tarsals. Including the green tree monitor, land iguanas, crocodiles, komodo dragons, and caiman lizards to name a few. I don't think it would be much of a stretch to include them in dinosaur reconstructions. Unless of course a dinosaur "mummy" was found indicating something else on a particular animal. I for one won't be overly shy in including them in my reconstructions. I've included links to examples of those I mentioned above.,

    1. you do understand you are talking about broadenned squamate scales, right? they are not homologous with bird scutes.

  12. if the retculate scales are "feathers arrested in early stages of development," why is it, then, that the "feathers came first" hypothesis is considered the more likely one? is there anything specific that prevents the reticulate scales from being considered the precursors of feathers rather than the other way around?