[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.
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.
Well, as is often the case, this post is a bit late to the party, despite starting early. Unless you have been living under a rock (or don’t care that much about dinosaurs), you have probably heard about the discovery of a small ornithischian from Siberia, Russia that apparently sports feathers as well as scales on its body. It’s a crazy half-and-half animal that has given many the green light for making all dinosaurs feathery.
As is often the case with these studies I am writing to urge caution against taking things too far, if just so there is some voice of dissent out there in an internet fully of trigger-happy feather reconstructions.
This post took a little longer to get together than I expected. Much like the first installment of this series, I found myself writing more and more. This time, though, rather than bother with breaking the post up into a bunch of smaller sections, I’ve decided to just dump the whole thing online at once.
Don’t worry, I’ve provided lots of pretty pictures to ease the eye strain. 🙂
While an in-depth look at Tianyulong confiusci‘s filaments (or as in-depth as one can get with just photos), has left me with doubts regarding their validity, one question still lingers.
If the filaments do prove to be genuine epidermal structures, then what does this mean for dinosaurs in general?
When this little ornithischian was announced, many in the paleo community (in particular the paleo-art community) seem to have used this little guy as a license to draw feathers on pretty much any dinosaur. After all, if protofeathers are found in ornithischians and saurischians, then it seems likely that they were a basal trait for dinosaurs in general. Some have even argued that the filaments alleged for Tianyulong, along with the protofeathers of maniraptorans, and the “fur” in pterosaurs, are all homologous structures; thus making a “furry” covering a primitive (plesiomorphic) trait for all of Dinosauria.
This is where we really need to start putting the brakes on. One only needs to do a cursory examination of any archosaur cladogram to see that there is a problem with this argument.
Though it is all too often forgotten, we have found the skin impressions from practically every major dinosaur group known to science. You know what these impressions show?
In practically every case, “skin” impressions from dinosaurs show them to have been scaly. Impressions from hadrosaurs (Sternberg, 1909, Anderson et al 1999), ceratopians (Brown 1917, Sternberg 1925), stegosaurs (Xing et al 2008, and photo on the left), ankylosaurs (Parks, 1924), sauropods – including embryos (Coria and Chiappe 2007), and most theropods (Abelisaurs [Czerkas & Czerkas 1997], Allosaurs [Pinegar et al 2003] and Tyrannosaurs [Currie et al 2003]) have all shown the presence of hexagonal, or tuberculate scales. Dinosaurs were a decidedly scaly bunch. (Proto)feathers were the exception, not the rule.
A common counter-argument to this has been that protofeathers could have been lost as animals got larger, or that protofeathers were an ontogenetic thing, with fuzzy babies going bald as they reached adulthood.
The essential problem with this argument is that scales are not equivalent to naked skin.
Scales, like hair and feathers, are a form of integument. Though they form as an infolding of the epidermis, they nonetheless lie on top of it. There are certain mutations in reptiles that will produce scaleless mutants (e.g. “silkback” dragons). These mutants retain their epidermis (which often looks very loose). The epidermis can also be clearly viewed between the scales of snakes while they are swallowing a large prey item. If dinosaurs really did lose protofeathers as they got larger, then one would expect to see patches of naked skin in between patchy feathers (much like what we see in extant pachyderms), but that’s not what we are seeing.
It is often pointed out that birds have both scales and feathers, thus making it possible for scales to occur in conjunction with feathers on dinosaurs.
However, this generalizes the relationship between scales and feathers. The fact is scales in birds do not occur because of an absence of feathers, but rather from active suppression of feather formation (Sawyer and Knapp, 2003). If one has ever plucked a chicken one might notice a distinct lack of scales on the most of the body. Despite the fact that feathers form along tracts in the skin, the areas between these tracts remain bare. Ostriches (Struthio camelus) provide another prime example of this.
Ostriches are large birds that, like most large animals living in tropical climates, have undergone a fair amount of insulation loss in order to avoid overheating. One need only look at the bare flanks, or neck of an ostrich to see that scales are nowhere to be found on these section. Scales only occur on the tarsometatarsal (ankle and toe) portion of the body. In fact there is a rather sharp demarcation where this occurs. This demarcation agrees well with embryonic studies of diapsids which show how integument formation occurs (Alibardi & Thompson 2001).
Feather ß-keratin proteins are likely homologous with scale ß-keratin. However they are also smaller than scale proteins (likely caused by a deletion to the scale ß- keratin gene [Gregg et al 1984]). Taken together all of this suggests an antagonistic relationship between scales and feathers. One that would determine integument placement based off of where one protein cascade ends, and another one begins.
To put it another way, the chances of a scaly dinosaur with a feathery mohawk, are extremely unlikely.
The ontogenetic argument seems even less likely, as it posits that dinosaurs lost one type of integument as hatchlings and then grew a completely different type as they reached adulthood. This would make dinosaurs unique among vertebrates in doing that.
To summarize then, scaly dinosaurs were not “naked” like elephants and rhinos. If we are to believe that a dinosaur group lost protofeathers as it evolved to be larger, then we must also assume that group then re-evolved scales in its place.
It is at this point where a cladogram comes in handy.
The following are three cladograms showing the possible evolution of filamentous integument in archosaurs. Each terminal group is one that we know the integument for (though not the exact member who’s picture I used). I’ve simplified things a bit with the coelurosaurs due to the nebulous nature of both Sinosauropteryx prima and the putative tyrannosauroid Dilong paradoxus. This should have little effect on the results as all these guys would do is add even more steps to the following situations. The general outcome remains unchanged.
The following are a few hypotheses that have been proposed over the last month for dinosaur integument evolution.
Hypothesis 1: The filaments seen in Tianyulong, Psittacosaurus, maniraptors, and pterosaurs are all homologous structures, thus making protofeathers the plesiomorphic trait for all of Dinosauria.
Take a look at our first cladogram. The blue dots indicate cases where a trait was lost, or reversed. In order for our first hypothesis to be true, then protofeathers would have to have been lost a total of 7 times! Also keep in mind what I mentioned previously. We are not just talking about protofeather loss, but also scale re-acquisition. That would also have to have occurred 7 times; making for a whopping 14 evolutionary steps!
Hypothesis 2: The filaments seen in Tianyulong, Psittacosaurus, maniraptors, and pterosaurs are merely analogous to each other. They represent yet another case of convergent evolution.
As the second cladogram shows; if this position is true, then protofeathers would have evolved a total of 4 different times. Once in the theropod line, once in pterosaurs, and twice in Ornithischians. That’s still a lot, but not nearly as many as in our first case.
Hypothesis 3: Protofeathers were the plesiomorphic trait for ornithodirans (pterosaurs and dinosaurs), but were lost at the base of Dinosauria, and subsequently reacquired by various dinosaur groups over time.
As one can see from cladogram 3 there, this situation results in a messy outcome. We see a single re-evolution in theropods, while Ornithischians show a helter-skelter pattern of filament reacquisition, and subsequent loss. The result is 1 case of evolution, 4 cases of filament loss as well as 4 cases of scale reversal, and 2 cases of filament re-evolution; making for a grand total of 11 steps.
Technically one could make the 3rd cladogram a bit different by having filamentous integument evolve twice within Ornithischia. This reduces the steps needed to 6, and makes for a cladogram very similar to cladogram 2.
A general rule of thumb for systematic paleontology, is to assume that evolution takes the least amount of steps possible (we assume Nature is generally lazy that way). As such, the evolutionary situation that produces the fewest “steps” is assumed to be the most likely situation. Nature doesn’t have to flow that way. There are cases out there where evolution might take a more complicated road, but in general this assumption that the simplest explanation is the most likely, tends to hold up.
So what does that say about our current situation?
Assuming that filamentous integument occurred a few times in ornithodiran evolution, results in a cladogram with substantially fewer steps (4). As such, it appears the most likely, or most parsimonious case.
Protofeathery integument could still be basal to Dinosaurs, and all those necessary reversals could still have occurred, but the road getting there seems unnecessarily complicated, and thus rather unlikely.
As it stands right now, it appears that if the filaments on Psittacosaurus and Tianyulong did belong to their respective owners, then they are a case of convergent evolution. Though generally frowned upon in systematics (mostly because it is a pain in the ass for phylogenetics), convergence is a rather common feature of evolution. For instance, in squamates alone the evolution of live birth has occurred a conservative 100 times (Shine 2005)!
So yeah, convergence happens; even for seemingly complicated things. That the filaments in these ornithischians, bear almost zero similarity to those of Sinosauropteryx and kin, further supports the hypothesis that they are an independent case of evolution.
There is another alternative that seems to rarely get mentioned. It is possibile that these filaments are actually scale derivatives. This would not be that surprising. Scales produce a wide variety of different ornamental structures in extant reptiles (from strange nose protuberances in certain iguanians, to flashy frills in agamids, and soft velvety skin in some geckos). In fact, the presence of the Psittacosaurus “quills” alongside scales, suggest that they are more likely to be a scaly derivative, than a feathery one.
Gonocephalus grandis, Rhacodactylus ciliatus, and Atheris hispida. Just some examples of scale diversity in extant reptiles.
What of the other major implication for basal “fuzz” in dinosaurs. Does this clinch the “dinosaurs were warm-blooded” argument?
Despite the wishes of some of the more vocal dino enthusiasts on the internet, this does not signal the death knell for bradymetabolic dinosaurs.
Both mammals and birds have an insulatory coat. From what we can gather, the role (or one of the roles) of this coat is to keep body temperature fairly constant. Therefore it is tempting to look at both feathery birds and fuzzy mammals and assume that a high metabolic rate (or automatic endothermy) must be associated with insulation.
However mammals and birds only represent two instances of insulation. As any statistician will tell you, two points make a line, not a pattern. What would help would be if there was at least one other group of critters that had insulation.
Well, it turns out that there are: Arthropods.
From the “woolly crustaceans” of the deep ocean, to bees and tarantulas, “hair” is fairly common among arthropods. This hair (deemed: setae) has a different embryological origin from mammalian hair, so it cannot be considered homologous.
So there is a third outgroup that shows filamentous coverings. Is it also associated with a constant body temperature and automatic endothermy?
In many species, the setae appear to function primarily as touch sensors; whether it be for the legs of a fly, or the body of a orb weaving spider. Still there are a few (moths, bees, certain beetles), that do use their hair for insulation. These animals are “functional endotherms.” That is to say that they use muscular power to generate heat internally. The difference between them and the classic “warm-blooded” mammals and birds, is that heat is generated solely by “skeletal” muscle, and can be turned off.
That insulation should not automatically equal “warm-bloodedness” has been recognized before. Previous authors (Schmidt-Nielson 1975, Withers 1992) have pointed out that while insulation does seem to lead to homeothermy, it does not associate so well with a high metabolism.
So then could we say that Tianyulong and the “feathered” theropods were using their insulation to maintain a stable body temperature.
If one is to use filaments for insulation, then they need to be spaced close enough that they will trap a layer of air between them and the skin. In mammals and birds this results in a notably fuzzy coat. Yet, sometimes this look can be deceiving. Consider polar bears. Despite their hairy look, polar bear fur offers very little insulatory benefits (Lavers 2000). The main use for the fur, seems to be to hide the black, sun absorbing skin underneath. Polar bears stay warm by maintaining a large layer of fat between their skin and the body core. The wide spacing of the hairs also allows them to quickly drain water from the body when the bears emerge from their icy swims (where insulation benefits of fur equal exactly zero). So if one is going to keep warm by being fuzzy, then that fuzz better be pretty thick.
For the protofeathered/feathered maniraptorans, the fuzz count appears high enough to allow for functional (possibly passive) homeothermy. This is not the case with Tianyulong. The filaments in T.confiusci are spaced too far apart to allow for much in the way of heat retention. These filaments must have been used for something else. Possibly as a means of defense by keeping attention focused on the tail, or (if backed by erector muscles) by making the animal look substantially bigger and more intimidating to a potential predator. They may have been used in a more passive sense by conferring camouflage to their owner. All are possible alternative uses for these filaments (ignoring, for now, the likelihood of these filaments being used for multiple purposes).
Besides all that, the Mesozoic is well known for being a time of high global temperatures. This doesn’t lend well to the assumption that filaments were evolved to keep their owners warm.
Now if they evolved to help keep heat out…
Anderson, B.G., Barrick, R.E., Droser, M.L., Stadtman, K.L. 1999. Hadrosaur Skin Impressions fom the Upper Cretaceous Neslen Formation, Book Cliffs, Utah: Morphology and Paleoenvironmental Context. Vertebrate Paleontology in Utah. David Gillette (ed). Utah Geo Survery. ISBN: 1557916349, 9781557916341 pps: 295-302.
Alibardi, L. and Thompson, M. 2001. Fine Structure of the Developing Epidermis in the Embryo of the American Alligator (Alligator mississippiensis, Crocodilia, Reptilia). J. Anat. Vol.198:265-282.
Brown, B. 1917. A Complete Skeleton of the Horned Dinosaur Monoclonius and Description of a Second Skeleton Showing Skin Impressions. Bul AMNH. Vol.37(10):281-306.
Coria, R.A. and Chiappe, L.M. 2007. Embryonic skin from Late Cretaceous Sauropods (Dinosauria) of Auca Mahuevo, Patagonia, Argentina. J. Paleo. Vol.81(6):1528-1532.
Currie, P.J., Badamgarav, D., Koppelhu, E.B. 2003. The First Late Cretaceous Footprints from the Nemegt Locality in the Gobi of Mongolia. Ichnos. Vol.10:1-12.
Czerkas, S. A., and S. J. Czerkas. 1997. The integument and life restoration of Carnotaurus. In D. L. Wolberg and G. D. Rosenberg (eds.), Dinofest International, Proceedings of the Symposium at Arizona State University, pp. 155?158. Philadelphia Academy of Natural Sciences, Philadelphia.
Gregg, K., Wilton, S.D., Parry, D.A., and Rogers, G.E. 1984. A Comparison of Genomic Coding Sequences for Feather and Scale Keratins: Structural and Evolutionary Implications. Embo J. Vol.3(1): 175-178.
Lavers, C. 2000. Why Elephants Have Big Ears: Understanding Pattersn of Life on Earth. St. Martins Press. NY. ISBN: 0312269022. pg 104.
Parks, WA. (1924). Dyoplosaurus acutosquameus, a new genus and species of armoured dinosaur; and notes on a skeleton of Prosaurolophus maximus. University of Toronto Studies, Geological Series 18, pp. 1-35
Pinegar, R.T., Loewen, M.A., Cloward, K.C., Hunter, R.J., Weege, C.J. 2003. A Juvenile Allosaur with Preserved Integument from the Basal Morrison Formation of Central Wyoming. JVP. vol.23(3):87A-88A.
Sawyer, R.H. and Knapp, L.W. 2003. Avian skin Development and the Evolutionary Origins of Feathers. J. Exp. Zool. (Mol Dev Evol). Vol.298B:57-72.
Schmidt-Nielson, K. 1975. Animal Physiology Adaptation and Environment. Cambridge University Press. Cambridge. ISBN: 0521570980, 978-0521570985. pg 669.
Shine, R., 2005. Life-History Evolution in Reptiles. Annu. Rev. Ecol. Evol. Syst. Vol.36:23-46.
Sternberg, C.H., 1909, A new Trachodon from the Laramie beds of Converse County, Wyoming. Science, v. 29, p. 753-754.
Sternberg, CM., 1925, Integument of Chasmosaurus belli: Canadian Field Naturalist, v.39, p. 108-110.