New Siberian ornithischian and the (over) feathering of dinosaurs…again.

Facebooktwitterredditpinterestlinkedintumblrmailby feather
Artist's impression of the fleshed out Kulinda specimen. Image by Andrey Atuchin
Artist’s impression of the fleshed out Kulinda specimen. Image by Andrey Atuchin

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.

Let’s start from the beginning.

This story has been cooking for quite some time. It all started in early 2013 during the abstract submission process for the Society of Vertebrate Paleontology meeting later that year. Pascal Godefroit and colleagues submitted an abstract on the fauna from a new fossil locality in Siberia, Russia known as the Kulinda valley. In the abstract (and title) Godefroit announced the discovery of a new neoornithischian that was alleged to preserve scales along the tail and legs, along with branched pennaceous (i.e., true) feathers on the body. Despite the SVP meeting being eight months away the buzz around this new taxon was already spreading like wildfire across the dinosaur paleo community. Many paleo-artists used this new taxon as further justification for feathering all dinosaurs.

The kicker, however, was that this new species did not see the light of day in 2013. During the SVP meeting, on the day of his talk, Godefroit had to cancel due to a leg injury he had sustained a few days earlier. Up until now there had been no details (not even a blurry picture) to indicate that this “feathery” ornithischian was real. There was further strangeness about this new locale. A few months prior to the abstract announcement news reports were coming out of Russia talking about a ‘Dinosaur Pompeii.’ These reports all referred to the Kulinda valley locality and its amazing preserved fauna, but they also talked about how this locality had been prospected in for the past two years. Yet despite all that time, not a single fossil had been described from there. News reports talked about the amazing preservation of scales and feathers in dinosaurs and other fauna. Thus it seems strange that such a high-impact discovery such as the Kulinda valley would take this long to have anything published. Having the talk about these fossils canceled at the last minute at SVP only further added suspicion to these finds. Our first look at the fossils of the Kulinda Valley (short of blurry news reports) finally came in early 2014 when Vladimir Alifanov wrote a brief communication that detailed the geological setting of the region. It is worth noting that even this short communication took two years to get published (it was originally accepted back in 2012). That is a remarkably long delay for any journal, and one that only further raised suspicions about what all was going on with these Russian fossils (more on this below).

The Kulinda valley is located in the Chernyshevskii district near the city of Chita in Siberia, Russia. According to Alifanov (2014) the strata in the Kulinda valley has been associated with the Ukureisk Formation. The exact age of this deposit is contentious and has ranged from Middle Jurassic to Early Cretaceous (Alifanov 2014). The current, middle ground, assessment of the Ukureisk Formation being Late Jurassic in age is based on the insect assemblage within the formation. As it currently stands the age of these rocks is in flux.  The paper does discuss the fauna of the valley, but does not go into much detail on anything.  The only figure to show some of the fossils is figure 2. This figure shows the jaw of an ornithischian (largely as a poorly preserved impression), the caudal vertebrae of a “theropod” with associated scales (apparently, though the quality of the photos does not allow for much discernment) and lastly a shot of the alleged bristles on an ornithischian. In the paper Alifanov refers to these structures as “bristle-like skin derivatives”. The figure only shows the “bristles”. No skeletal association can be discerned. This was done on purpose, apparently to tease at the find more than anything else. As I mentioned earlier, the photos are extremely poor quality (approximately 538 x 401 pixels, and in black and white), so there is not much that can be ascertained from them.

However now, one year after the SVP abstract, we have the release of not one, but three papers on this “feathery” creature, as well as three different names for it.

Saveliev, S.V., Alifanov, V.R. 2014. A New Type of Skin Derivatives in Ornithischian Dinosaurs from the Late Jurassic of Transbaikalia (Russia). Doklady Biol. Sci. Vol. 456:182–184.

Alifanov, V.R., Saveliev, S.V. 2014. Two New Ornithischian Dinosaurs (Hypsilophodontia, Ornithopoda) from the Late Jurassic of Russia. Paleontological Journal (Russian Edition). Vol. 4: 72–82.

Godefroit, P., Sinitsa, S.M., Dhouailly, D., Bolotsky, Y.L., Sizov, A.V., McNamara, M.E., Benton, M.J., Spagna, P. 2014. A Jurassic Ornithischian Dinosaur from Siberia with both Feathers and Scales. Science. Vol. 345(6195):451–455.

What’s my name again?

Faux NMMNHS bulletin
Remember this mess?

Why three papers? Well it turns out that this Kulinda valley ornithischian was a hot-ticket item. Astute readers no doubt noticed that the first two papers have a completely different authorship from the third paper. To explain the reason behind it I’ll turn to a quote from Pascal Godefroit in a Facebook post (turned DML post) from a few weeks back:

These are based on specimens that Alifanov stole in Chita, and that are illegally housed in PIN in Moscow, but belonging in fact to the Institute of Natural Resources, Ecology and Cryology (Chita).

So it would appear that we have another Aetogate on our hands here. Alifanov and Sergei Saveliev seem to have intentionally claim jumped Godefroit et al. with regards to this specimen. Further confusing matters is the fact that Alifanov and Saveliev described two different ornithischian taxa where Godefroit et al. only saw one. So this new ornithischian is either Kulindadromeus zabaikalicus, Kulindapteryx ukureicaDaurosaurus olovus, or some combination thereof. Following the standard rules of taxonomic nomenclature, the first published work out the gate gets priority. This means that the new ornithischian would technically be K. ukureica and D. olovus. However, if the analysis is bad (i.e., the material is not diagnostic) then it is possible that the names could be sunk as nomina dubia. This could free up K. zabaikalicus for use, but the fact that it has come out weeks after the description by Alifanov and Saveliev means that it might get sunk as a junior synonym of the first two names and thus could get sunk with them both. This has the potential to be a huge taxonomic nightmare and will likely require a formal petition to the International Coalition for Zoological Nomenclature (ICZN) to fix.

Well that, or everyone will just choose the name they prefer and just ignore the others, basically working it out via a de facto consensus. Either way these specimens are now viewed as tainted, which could greatly affect their research potential (just look at how much has been done with the allegedly filamented Psittacosaurus, which was also obtained via illegal means).

For the purpose of this blog post I am going to refer to these specimens as the Kulinda valley ornithischian.

 

Bristles, feathers and scales, oh my!

Though the taxonomic validity (biological bookkeeping) is likely to remain contentious for some time, the actual structures on this animal can be discussed much more easily. Our first real information on these structures came from Alifanov and Saveliev 2014. Here, in between discussing the taxonomic affinities of the animals in question, the authors set about describing the integument. Since this came from a journal that was published in Russian, the following translation by Dave Černý (posted here) may not be 100% accurate to the original text. Still it is a hell of a lot better than what Google translate provides, and the gist can be fully grasped.

[NOTE: An English version of Alifanov and Saveliev has since been published. Comparing the quoted material below with the English translation reveals little differences. The only major difference was the changing of ‘scutes’ into ‘plates’. As such I have left Dave’s transcription below and added the English-language version of Alifanov and Saveliev to the references section.]

Three different interpretations of the integument distribution in the Kulinda ornithischian. Skeletals adapted from figure 4 of Alifanov and Saveliev 2014 (top two) and figure 1 of Godefroit et al. 2014 (bottom).
Three different interpretations of the integument distribution in the Kulinda ornithischian. Skeletals adapted from figure 4 of Alifanov and Saveliev 2014 (top two) and figure 1 of Godefroit et al. 2014 (bottom). Click to enlarge.

From Alifanov and Saveliev 2014 (translated. bold type mine):

The impressions of integument in PIN 5435/57 are located close to the humerus and are concentrated in front of it (mostly at its proximal end) and behind it, where they are represented by [skin] derivatives of two types. One of them takes the form of oval scutes or scales, the other one [consists of] bristle-like structures. In this specimen (Plate XI, Fig. 4), around the posterior margin of the humerus, the “bristles” lie atop the scutes. However, some of the “bristles” appear to be extensions of the scutes. If so, the “bristles” can be either the result of scale splitting or the consequence of their differential growth. Judging by the arrangement of the areas [of bristles] in front of and behind the humerus, the described type of integument evenly covered the surface of the forelimb, and perhaps of the whole body. Interestingly, with regard to the length of the bristles, this fossil differs from the one that is described below.

In PIN 5435/56, the existence of which has played a role in the choice of the name for Kulindapteryx[Kulinda wing]ukureica gen. et sp. nov., long bristle-like structures are present next to the bones of the forelimb (the humerus is preserved on the main slab and the radius on the counterslab), forming an extensive halo in the matrix similar to wing impressions of fossil birds.

The preservation of the bones of the last specimen makes even an approximate systematic placement impossible. However, considering that two length variants of bristle-like structures on the same bones correspond to two species of hypsilophodontids, it is not ruled out that in Kulinda, different species of hypsilophodontids had “bristles” of different length. If so, it would be in agreement with the pattern of predominance of Daurosaurus olovus gen. et sp. nov. remains in the locality, which we have established, and judging from the distribution of specimens with impressions of bristles of different length, it is possible to tentatively assume that it was the latter species that had the short-bristled type of scales. In that case, it remains to infer the presence of the long-bristled type of integumental appendages for Kulindapteryx ukureica gen. et sp. nov.

The authors appear to allude to knowledge of many more specimens from the Kulinda valley. As mentioned above, their decision to separate these two skeletons into two different genera is questionable, as is their assignment of these taxa to Hypsilophodontidae (apparently the diagnostic characters for anything short of Neornithischia, just aren’t preserved). It’s also worth noting that the authors refer to slabs and counterslabs. This tells us that we are looking at flat, “pancake” fossils similar to the types of fossils that pervade Liaoning, China.

However the most interesting point of this entire description comes from the association of the scales with the bristles. As I have mentioned previously on this site (extensively), our knowledge of feather evolutionary development strongly suggests that feathers and scales cannot coexist. At the very least it appears to be extremely difficult to do so. The likely re-evolution of scales in birds (Dhouailly 2009) further suggests that the evolution of feathers was a punctuated event that was likely caused by changes in the developmental mechanisms that form scales. So the presence of bristles that “lie atop the scutes [plates]” strongly suggests that what we are looking at in these ornithischians are not protofeathers but some entirely different type of integument. Unfortunately the authors images of these specimens are just about as piss poor as the previous paper (Alifanov 2014), making it extremely difficult to determine what these structures look like (though plate XI-4 does have some promise). The authors do provide a skeletal drawing of the specimens indicating that the bristly structures covered the entire body (contra Godefroit et al. 2014. However, look at how much was preserved vs. interpretation [see above image]). The exact structure of these bristles is further elaborated on in Saveliev and Alifanov 2014.

Integumentary remains of ornithischian dinosaurs are usually represented by bristle-like outgrowths, which are arranged on the matrix surfaces either chaotically or in bunches or regular groups of varying density. Some samples show the basal portion of the “bristles” in the form of a small plate with a rounded proximal end. One plate can give rise to several outgrowths of various lengths. We propose to call these integumentary structures “bristle-like scales.” No such structures have previously been found in association with ornithopod or ornithischian remains….

…. Samples with integumentary impressions most frequently have scale with three or four “bristles.” The next most common are scales with five bristles, in which the central “bristle” is usually longer than the four lateral ones. The six, seven, and eight “bristle” scales are less common… In some samples, long “bristles” are observed as a dense entangled pattern. — Saveliev and Alifanov 2014

The authors provide a series of pictures to illustrate their description of the bristle-like scales. Unfortunately there are no pictures of the structures on the associated specimens (likely due to the illicit nature of their acquisition). Nonetheless the anatomical description of the bristle formation is extremely interesting and is suggestive of a new scale-type. Unlike feathers, these structures actually appear to form by a “fraying” and differential growth of the distal end of the scale. That this “fraying” is so varied may explain why the original Godefroit et al. abstract referred to these as “compound feather-like structures“.

Saveliev and Alifanov go further in their description, talking about the pigmentation associated with these structures and what that may mean for feather development.  In particular:

Usually, the scales are pigmented similarly and relatively uniformly along the entire length. Portions of preserved dermal fields were formed only from compactly arranged dark and short-bristle scales. Dark colored “bristles” occur in isolation or in bunches of varying appearances and size. The uniform dark color is characteristic of scales possessing five or fewer “bristles.” Less commonly found are weakly segmented “bristles” connected with the dark-colored basal ends of the scales. Initial parts of the bristle-like structures in such scales are also dark-colored (Fig. 1). A two color pattern is observed mainly in six-to-eight-bristle scales. If the pigment border is taken to be a marker separating the introdermal and extrodermal zones of scales, it appears that some scales and the bases of the bristles are submerged. Thus, bunches of “bristles” extending from the skin could be connected to each other at their bases inside the derma. It is possible that a similar structure could occur on the dorsal end of the tail of the bipedal horned dinosaur Psittacosaurus, judging from a skeleton with skin impressions from the Lower Cretaceous of China.

It’s an interesting interpretations, but there isn’t really much reason to suppose this darker pigmentation should represent the epidermal outline. If so that would make these scales dermal structures more akin to the scales of fish and amphibians. These structures are strange enough on their own without having to make them any stranger. This is not the first time that Saveliev and Alifanov have made questionable interpretations of data. A similarly strange interpretation of ornithopod neuroanatomy was made by these authors two years ago wherein they interpreted fossae in the endocast of Amurosaurus riabinini as being bulbs for the vomeronasal organ (Saveliev et al. 2012) despite evidence that strongly supports loss of this structure in Archosauria (Senter 2002). I suspect that part of the reason for the authors argument for a dermal structure to these scales comes from the part later in the paper where they attempt to homologize these structures with the filaments seen in advanced coelurosaurs. However, it is done in a very slipshod manner, as evidenced by a total lack of citations in regards to the evolutionary development of feathers. The paper itself is extremely light on citations anyway (six total), further suggesting that these authors were attempting to beat Goldefroit et al. to the punch.  So no, I don’t buy their case for dermal scales with epidermal bristles coming off them.

Now Godefroit et al. have, by far, the most in-depth study on these specimens (even if most of it resides in the supplemental section). The Kulinda specimen description is based on six different skulls along with hundreds of disarticulated skeletons. The species is known from two monospecific bonebeds. That they were unearthed by paleontologists and not bought off a seedy black market, tells me that we are dealing with the real deal and not some potential forgery. Godefroit et al. argue that the preserved bristles are feathers. Not just filaments, but full-on feathers. Such an interpretation seems rather bold given that their description of the bristles is in near complete agreement with that of Saveliev and Alifanov. The one notable difference is that Godefroit et al. are less willing to call the plate-like structures that the filaments are coming off of, scales.

Kulindadromeus also shows compound, nonshafted integumentary structures along the humerus and femur… These occur as groups of six or seven filaments that converge proximally and arise from the central regions of a basal plate….Whether the basal plates represent modified scales or calamus-like structures remains unclear and requires further investigation.

Interpretations of these compound filamentous structures. Images adapted from plate XI-4 of Alifanov and Saveliev 2014 (top left), figure 1 of Saveliev and Alifanov 2014 (top right), and figure S9-B, C of Godefroit et al. 2014.
Interpretations of these compound filamentous structures. Images adapted from plate XI-4 of Alifanov and Saveliev 2014 (top left, red outline of forelimb bones for ease of contrast), figure 1 of Saveliev and Alifanov 2014 (top right), and figure S9-B, C of Godefroit et al. 2014.

The authors also discuss a third filamentous structure observed on the Kulinda ornithischian. This was a series of ribbon-like structures located near the proximal tibia. They are longer than the monofilaments and the compound filaments. They are also thicker, resembling leaves, almost (Godefroit et al. 2014 figure 3H). Interestingly, the only other integumentary impression like this are the pseudofeathers of the archosauromorph Longisquama insignis. However, unlike L. insignis, these structures laid close to the bone surface, and thus, may not even be integumentary, but some type of subdermal structure.

Half and half?

Along with these unique filamentous structures is their strange distribution. Monofilaments are found along the head, neck and thorax. The femur and humerus house the strange compound structures. The forearm, leg (tibia and fibula) and tail, meanwhile, contain scales. The scales range from tuberculate scales reminiscent of the scales seen in hadrosaurs (Godefroit et al. argue that they look more like bird scales, but as hexagonal structures I just don’t see the resemblance) to slightly overlapping scales on the tail that look similar to the belly scutes of crocodylians. Keels on the dorsal scales of the tail, along with their different colouration, makes them seem like the dorsal scutes of some turtles. So we have a critter that seems to have a filamentous core with scaly extremities. The compound structures look like they may have been used for display, but their location and orientation make it a bit difficult to tell for certain. Interestingly, as these compound sections approach the distal humerus they appear to grade into the scales of the forearm (Godefroit et al. 2014 figure 4A). One of the author’s misgivings about calling these “basal plates” scales was their rather spread out nature. They neither overlap nor abut each other. However, as these compound structures approach the forearm this splaying starts disappearing and these “basal plates” end up looking very similar to the scales on the forearm. That said, once one reaches the humeroulnar joint its not really possible to distinguish individual scales anymore.

Some gecko lineages such as this bronze gecko (Ailuronyx tachyscopaeus) are capable of "wiggling out of their skin" when attacked by a predator. Was the Kulinda ornithischian doing something similar? Photo by Henrik Bringsøe
Some gecko lineages such as this bronze gecko (Ailuronyx tachyscopaeus) are capable of “wiggling out of their skin” when attacked by a predator. Was the Kulinda ornithischian doing something similar? Photo by Henrik Bringsøe.

As far as I know such a strange arrangement of integument is unheard of in the modern world, which makes it hard to determine what the point of such strange structures was.  The typical go to answer of ‘insulation’ doesn’t really make much sense. Yes, the animals were rather small at their (presumed) adult size, and thus would benefit from some kind of insulation better than larger dinosaurs, but then so was the scaly Juravenator starki, as were the babies of all known scaly dinosaurs. Looking at the life reconstructions of the animal one can see a definite mammal-centric vibe to the reconstructions, with tucked in forelimbs and a thin tail. However, as previous anatomical studies (Allen et al. 2009, Persons and Currie 2011) have indicated, dinosaurs were ‘rear-wheel drive’ animals (Hotton 1994) that received most of their leg retracting power from the large caudofemoralis muscles located on the tail. This would have given the Kulinda ornithischian a larger, thicker tail. One would suspect that, if the primary function of these structures was body-heat maintenance, then at least the proximal part of that tail would have been covered in fuzz. Yet scales seem to take over immediately caudal to the hips. Another problem comes from the inferred paleoenvironment for the Kulinda valley. It was believed to have been a swamp at the time. That’s hardly a place where an animal has to worry about keeping warm. The rather diffuse coverings of filaments would also suggest that body-heat maintenance may not have been the primary function of these structures. Were they anti-predator defenses? Perhaps, but it would seem strange that the extremities were not also covered in filaments too, as these are more likely to meet the business end of a predator’s mouth. Then again, having a fuzzy core could reduce the damage inflicted from getting a direct hit by a predator. Perhaps these structures were used in a similar vein to that of some gecko species in which the skin sloughs off very easily, allowing for a quick escape. Of course all of these are just speculations based on no evidence. Without detailed information of the ultrastructure of these filaments there really isn’t much that can be said about their function right now.

What the Kulinda ornithischian means for Dinosauria

Last year it only took an abstract to get many in the feathered-dinosaur movement (the self-described ‘Feather Nazis’) overly excited and insistent on covering every dinosaur species with a thick layer of plumage (the so called ‘enfluffening‘). Now that the beast is actually out what are its implications for the rest of Dinosauria?

If these filaments were homologous and thus ancestral to Dinosauria. Blue circles indicate how often scales would have to re-evolve. Click to enlarge.

If filaments evolved convergently in Dinosauria. Red dots indicate the amount of times filaments would have to evolve. Click to enlarge.

First let’s start with where the Kulinda specimens show up in the family tree of Dinosauria.

As is often the case, the answer is not that clear cut. Alifanov and Saveliev argued for placement within Hypsilophodontidae. However they provided no phylogenetic analysis to back this up. Godefroit et al. did better with their parsimony analysis. Their results suggested that the Kulinda specimens were basal neornithischians most closely related to Cerapoda (hadrosaurs and ceratopsians). However, the authors pointed out that their most parsimonious trees still had very little support for their topology. In fact Bremer support and bootstrap values were so low that the authors didn’t even report them for the node that the Kulinda specimen was in (Bootstrap proportions lower than 50 and Bremer decay values lower than 2 are not indicated — Godefroit et al. 2014 supplement). Although the authors nested these specimens deeper in the phylogenetic tree of Ornithischians than Alifanov and Savaliev did, the statistical support for this position was so low that we really can’t say for certain where these critters belong.

What that means is that this critter is going to bounce around for awhile until a better description of this and other ornithischians is put out. Judging from the shape of its skull (especially the large palpebrals) it is tempting to stick it either in Hypsilophodontidae or Heterdontosauridae, near Tianyulong confuciusi. The skulls are a bit different in their dentition, though the enlarged dentary canine of T. confuciusi is not seen in the Kulinda specimen, that portion of the skull is also not preserved. Judging from the description, it looks like the Kulinda specimen shares many traits in common with Heterodontosaurus and Hypsilophodon, but also plenty of things that are different. This paper illustrates just how little we know about ornithischian relationships in comparison to theropods. Still, if I were to predict placement, I would not be surprised if the Kulinda specimen wound up grouping closely with Tianyulong confuciusi, either as a heterodontosaurid, hypsilophodontid, or as something exclusive of the two.

 

On to the $64,000 question: Does the Kulinda specimen favour a “feathers on everything” approach to Dinosauria? Looking at the above two images we can see that the addition of the Kulinda specimens doesn’t really change the distribution of filaments within dinosaurs. Most dinosaurs are still scaly and it is still most parsimonious to assume that these structures evolved separately in ornithischians and advanced coelurosaurs.

Unfortunately neither group of authors took this position and instead opted for pushing “feathers” back to the base of Dinosauria, if not Ornithodira. Both groups argue for either homology of these filaments with feathers, or for just calling them feathers outright.

…bristle-bearing scales should be considered an important stage of skin evolution in dinosaurs, showing the transition from a simple scale to a birdlike feather….The above interpretation makes it possible for Ornithischia to be included in the concept of “Feathered dinosaurs.” — Saveliev and Alifanov 2014

The presence of both simple and compound filamentous structures in Kulindadromeus…supports the hypothesis that the integumentary structures in Ornithischia, already described in Psittacosaurus and Tianyulong, could be homologous to the”protofeathers” in non-avian theropods. In any case, it indicates that those protofeather-like structures were probably widespread in Dinosauria, possibly even in the earliest members of the clade. — Godefroit et al. 2014

Both groups appear to have ignored the extensive evolutionary developmental studies that have been performed on feather formation. We now have a pretty good idea of how feathers formed. In no case do feathers form from the distal portions of completed scales, which is what the Kulinda specimens seem to be showing.  At the very least, they don’t form from a “basal plate.” Feathers form during the formation of the embryonic epidermis, in particular the periderm and subperiderm. At this stage scales and feathers are the same thing (and don’t look like much). Changes in developmental timing, determined by a host of biochemical factors such as bone morphogenetic protein (BMP) and sonic hedgehog (SHH)  determine whether a feather will form or if a scale will form and then slough off the embryonic periderm (Sawyer et al. 2003, Alibardi et al. 2006). Dermal-epidermal recombination studies, in which the epidermis of one integument forming region is grafted to the dermis of another, has succeeded in achieving  “feathery scales” (Rawles 1963). However these should not be looked at as evidence that structures could, or did exist in the past, as they involved heavy tissue recombination. The purpose of these studies was to determine which portion of the skin (dermis or epidermis) was responsible for inducing which type of integument. Similar studies (Kato 1969, Dhouailly 1973) have succeeded in having mammalian hair grow on the dermis of chickens or feathers grow from the ectoderm of the extraembryonic fluids (chorion). Studies such as these reveal the mechanisms behind integument formation; they don’t recover lost phenotypes.

It is worth noting that one of the authors on the Godefroit et al. paper was Danielle Dhouailly, an evo-devo researcher who has done extensive work on the origin and formation of feathers. Unfortunately it seems as if Dhouailly’s contribution to the paper was mostly as a consultant for feather formation in birds rather than as an analyst of the Kulinda specimen’s structures. Thus the hope that this would be the first ornithischian to have its integument vetted have been partially dashed. I write ‘partially’ because the supplementary material does provide a couple of tests for alternate origins of these structures. The authors looked at the alternative interpretations for these structures as collagenous rot or coincidental plant material. They dismissed both, and provided good reasons for their rejections. For that I give the authors credit, and I accept their interpretation that these structures were part of the individuals.

The anatomy of a feather. Image from:
The anatomy of a feather. Image from the Arizona State University School of Life Sciences

However I differ pretty strongly with their repeated use of ‘feather’ in the paper and the supplement. Feathers are hierarchically branched structures arising from the epidermis (Prum 1999, Chuong et al. 2003). They are composed of a matrix of intracellular keratin and are anatomically composed of a follicle, calamus (base), rachis (central shaft), barbs (branches off the rachis) and barbules (branches off the barbs). That is a feather. To date pennaraptorans are the only animals known to exhibit feathers. The structures on Sinosauropteryx prima, Beipiaosaurus inexpectus, Dilong paradoxus, and Yutyrannus huali are not feathers. These are filamentous structures (epithelial appendages [Chuong et al. 2003]) that may be ancestral to feathers. They are commonly referred to as protofeathers (interestingly Godefroit et al. put protofeathers in quotes, but don’t do the same for feathers). The alleged structures on that one Psittacosaurus individual, as well as the structures on Tianyulong confuciusi, and the pycnofibres on pterosaurs are basically simple shafts, like hairs or arthropod setae. All of these structures match Prum’s (1999) predicted stage 1 protofeather, highlighting the degree of convergence that basic filaments have.

The simple filaments found on the Kulinda specimens similarly fall in the “stage 1 protofeather” / basic filament category. The more complex structures also found on the specimens don’t fit Prum’s model, as they are coming off these basal plates/scutes and they apparently lack rachides.

…except for our monofilaments (which correspond well to Type 1 in Xu et al. [2010. Note this is Prum’s stage 1 feather]), we could not assign with confidence the other two feather morphotypes in Kulindadromeus to categories described by Prum et al. [1999]. or Xu et al. [2010] — Godefroit et al. 2014 Supplementary information.

 

So I find it quite strange and disheartening that Godefroit et al.—despite being fairly objective in their supplementary material—go completely gung-ho in calling these structures feathers. Instead of talking about a really neat and new integumentary structure for dinosaurs, or at least pulling a Xing Xu and saying: “they could be protofeathers, or they could not “, we get:

…diverse epidermal appendages, including grouped filaments that we interpret as avianlike feathers.

This is very reminiscent of the initial description fo the carcharodontosaur Concavenator corvatus (Ortega et al. 2010) in which the authors found raised bumps along the ulna that they interpreted solely as quill knobs. No alternate explanation or detailed comparison with the quill knobs of birds were given. Now, four years later, there are many who believe that these bumps agree better with intermuscular septa than with quill knobs (Foth et al. 2014). Nonetheless, it’s not hard to find C. corvatus still grouped in with examples of “feathered dinosaurs”.

The potential fuzziness of archosaurs

Examples of non-feathery filamentous (or near so) integument in diapsids. The adhesive setae of geckos (left), the dorsal spines of iguanas (center) and the "beard" of turkeys. Gecko image by Danielle Whittaker. Iguana image by: Christian Ziegler. Turkey "beard" image by: Arlene Koziol
Examples of non-feathery filamentous (or near so) integument in diapsids. The adhesive setae of geckos (left), the dorsal spines of iguanas (center) and the “beard” of turkeys. Gecko image by Danielle Whittaker. Iguana image by: Christian Ziegler. Turkey “beard” image by: Arlene Koziol.

Pushing feathers back to the base of Dinosauria or Ornithischia seems to be the trend for many in the dinosaur paleontological community these days. However, as discussed above, this requires that many criteria be met in order for this statement to be true. Not the least of which is that we need evidence that the structures seen in all these filamented dinosaurs are actually homologous. Following Patterson (1982), this means that these structures need to pass the three tests of homology:

  1. Similarity — is the anatomy the same (shape, location to other structures, development)?
  2. Congruence — is it monophyletic?
  3. Conjunction — is a structure found to the exclusion of its alleged homologue?

These tests are meant to be done in succession. So similarity (primary homology) would be tested first followed by congruence (secondary homology) and conjunction (tertiary homology). To date there has been no test of primary homology on any filamented dinosaur specimen, save perhaps Sinosauropteryx prima and Caudipteryx zhoui.  In regards to the Kulinda ornithischian, a cursory look at the description indicates that only the monofilaments share any similarity with the structures seen in advanced coelurosaurs and Tianyulong confuciusi. They also match mammalian hair and insect setae, at least on a gross level, so that doesn’t tell us very much. Evidence of a follicular structure would go a long way to showing homology, but that requires an analysis of the ultrastructure, which has yet to happen. So on the outset the first test for homology (primary homology) isn’t looking good. Theoretically one need not test the other two if the first test fails, but if we were to generously assume primary homology of these structures, we can still test for congruence. That is exactly what Paul Barrett did last year at SVP (Barrett 2013). He compiled a list of all known dinosaur integumentary structures. He then used both a parsimony approach and a maximum likelihood approach to assess the ancestral character state for Dinosauria. Barrett’s analysis indicated that there is little support for a deep nesting of feathers within Dinosauria. This analysis came out a few months later in a news release in the journal Nature. Though a full-on paper would have been preferred this is, nonetheless, the only test for secondary homology of feathers within Dinosauria, and it was yet another test that feathers failed. As it currently stands there is still no support for homologizing the filaments in ornithischians with the filaments in advanced coelurosaurs, or the filaments in pterosaurs. To quote Barrett:

“I’d go so far as to say that all dinosaurs had some sort of genetic trait that made it easy for their skin to sprout filaments, quills and even feathers”

Many folks seem to have trouble swallowing the concept of filaments arising multiple times during archosaur evolution. It is important to keep in mind that we are talking about 185 million years, or approximately 3 times the span of time from dinosaur extinction to the rise of humans. That is a long time for evolution to tinker around. Roger Sawyer and his colleagues were the first to suggest that archosaur integument had an inherent capacity to form multiple epidermal appendages, due to their discovery of unique embryonic epidermal layers in developing birds and alligators (Sawyer and Knapp 2003, Sawyer et al. 2005). So it’s not like Barrett’s proposal was unique. We have extant examples of simple filaments arising on at least three different occasions in extant diapsids (see the image above), so there is already precedent for convergent evolution of filaments within Diapsida. In many cases we are just talking about the evolution of simple filaments. Only pennaraptorans seem to have elaborated beyond simple filaments, and that’s including the Kulinda specimens, which may have weird structures but they are still fairly simple compared to protofeathers and actual feathers.

For comparison keep in mind that Squamata has evolved viviparity a conservative 100 times (Shine 2005)! In terms of complicated developmental processes, going from an external egg to an internally supported baby (including the evolution of a placenta) is way more difficult than sprouting a long, thin filament.

So in the end the Kulinda specimen offers us new insight into the diversity of epidermal derivatives found in dinosaurs, but based on our current knowledge it doesn’t actually do anything to “fluff up” the dinosaur family tree anymore than it already is.

~ Jura

References

Alibardi, L., Knapp, L.W., Sawyer, R.H. 2006. Beta-Keratin Localization in Developing Alligator Scales and Feathers in Relation  to the Development and Evolution of Feathers. J. Submicro. Cyt. Path. Vol. 38(2–3):175–192.
Alifanov, V.R. 2014. The Discovery of Late Jurassic Dinosaurs in Russia. Doklady Earth Sciences. Vol. 455(2):365–367.
Alifanov, V.R., Saveliev, S.V. 2014. Two New Ornithischian Dinosaurs (Hypsilophodontia, Ornithopoda) from the Late Jurassic of Russia. Paleontological Journal. Vol. 48(4):414–425.
Alifanov, V.R., Saveliev, S.V. 2014. Two New Ornithischian Dinosaurs (Hypsilophodontia, Ornithopoda) from the Late Jurassic of Russia. Paleontological Journal (Russian Edition). Vol. 4: 72–82.
Allen, V., Paxton, H., Hutchinson, J.R. 2009. Variation in Center of Mass Estimates for Extant Sauropsids and its Importance for Reconstructing Inertial Properties of Extinct Archosaurs. Anat. Rec. Vol. 292:1442–1461.
Barrett, P. 2013. Dinosaur Integument: What do we Really Know? JVP.  Programs and Abstracts:82A.
Chuong, C-M., Wu, P., Zhang, F-C., Xu, X., Yu M., Widelitz, R.B., Jiang, T-X., Hou, L. 2003. Adaptation to the Sky: Defining the Feather with Integument Fossils from Mesozoic China and Experimental Evidence from Molecular Laboratories. J. Exp. Zool. (Mol. Dev. Evol.). Vol. 298B:42–56.
Dhouailly, D. 2009. A New Scenario for the Evolutionary Origin of Hair, Feather, and Avian Scales. J. Anat. Vol.214:587-606.
Dhouailly, D. 1973. Dermo-Epidermal Interactions between Birds and Mammals: Differentiation of Cutaneous Appendages. J. Embryol. Exp. Morph. Vol. 30(3):587–603.
Foth, C., Tischlinger, H., Rauhut, O.W.M. 2014. New Specimen of Archaeopteryx Provides Insights into the Evolution of Pennaceous Feathers. Nature. Vol. 511 Supplement:1–83.
Hotton III, N.. 1994. Why Dinosaurs were Not Mammals and Vice Versa. Dino Fest. Special Publication of the Paleontological Society. Vol. 7:39–59.
Kato, Y. 1969. Epithelial Metaplasia Induced on Extraembryonic Membranes. I. Induction of Epidermis from Chick Chorionic Epithelium. J. Exp. Zool. Vol. 170:229–252.
Ortega, F., Escaso, F., Sanz, J.L. 2010. A Bizarre, Humped Carcharodontosauria (Theropoda) from the Lower Cretaceous of Spain. Nature. Vol. 467:203–206.
Patterson, C. 1982. Morphological Characters and Homology. in Joysey, K.A., Friday, A.E. (eds): Problems of Phylogenetic Reconstruction. New York: Academic Press, pp. 21–74.
Persons IV, W.S., Currie, P.J. 2011. The Tail of Tyrannosaurus: Reassessing the Size and Locomotive Importance of the M. caudofemoralis in Non-Avian Theropods. Anat. Rec. Vol. 294:119–131.
Prum. R.O. 1999. Development and Evolutionary Origin of Feathers. J. Exp. Biol. (Mol. Dev. Evol.). Vol. 285:291–306.
Rawles, M.E. 1963. Tissue Interactions in Scale and Feather Development as Studied in Dermal-Epidermal Recombinations. J. Embryol. Exp. Morph. Vol.11(4):765–789.
Sawyer, R.H., Knapp, L.W. 2003. Avian Skin Development and the Evolutionary Origin of Feathers. J.Exp.Zool.(Mol.Dev.Evol) Vol.298B:57-72.
Sawyer, R.H., Rogers, L., Washington, L., Glenn, T.C., Knapp, L.W. 2005. Evolutionary Origin of the Feather Epidermis. Dev. Dyn. Vol. 232:256–267.
Sawyer, R.H., Salvatore, B.A., Potylicki, T-T.F., French, J.O., Glenn, T.C., Knapp, L.W. 2003. Oriign of FEathers: Feather Beta Keratins are Expressed in Discrete Epidermal Cell Populations of Embryonic Scutate Scales. J. Exp. Zool. (Mol. Dev. Evol.). Vol.295B:12–24.
Saveliev, S.V., Alifanov, V.R. 2014. A New Type of Skin Derivatives in Ornithischian Dinosaurs from the Late Jurassic of Transbaikalia (Russia). Doklady Biol. Sci. Vol. 456:182–184.
Saveliev, S.V., Alifanov, V.R., Bolotsky, Y.L. 2012. Brain Anatomy of Amurosaurus riabinini and Some Neurobiological Peculiarities of Duck-Billed Dinosaurs. Paleo. J.  Vol. 46(1):79–91.
Senter, P. 2002. Lack of Pheromonal Sense in Phytosaurs and Other Archosaurs, and its Implications for Reproductive Communication. Paleobiology. Vol. 28(4):544–550.
Shine, R., 2005. Life-History Evolution in Reptiles. Annu. Rev. Ecol. Evol. Syst. Vol.36:23-46.
Xu, X., Zheng, X., You, H. 2010. Exceptional Dinosaur Fossils Show Ontogenetic Development of Early Feathers. Nature. Vol. 464:1338–1341.
Facebooktwitterredditpinterestlinkedintumblrmailby feather

13 Responses to New Siberian ornithischian and the (over) feathering of dinosaurs…again.

  1. When I first got wind of this find, I had assumed that the Kulinda ornithischian’s peculiar integument represented a specialized adaptation for the environment of Jurassic Siberia. I don’t know exactly what that region’s paleo-climate would have been like, but I’m guessing it was cool by Mesozoic standards at least. There actually are wetlands with cooler climates.

    But yeah, this is another welcome counterpoint to all the over-feathering enthusiasm.

  2. First of all, thanks for the mention — I’m glad you found the translation useful.

    As I have mentioned previously on this site (extensively), our knowledge of feather evolutionary development strongly suggests that feathers and scales cannot coexist.

    You have mentioned it repeatedly, but I don’t think you have ever provided a citation for it. The papers you cite suggest that feathers share many similarities with scales (including nonavian scales) in the placode stage and both evolved through a modification of the same ancestral developmental pathways (e.g., Sawyer et al. 2003; Dhouailly 2009). They also show there is an antagonistic relationship between feathers and all the various types of bird scales (scuta, scutella, reticula), which are formed by active feather suppression. However, the antagonism obviously doesn’t preclude their coexistence, since they do coexist on the body of every single living bird. The suppression is local, not global. Correct me if I’m wrong, but none of the many evo-devo papers you have cited in your articles on this topic actually supports your conclusion that the presence of feathers absolutely rules out the presence of scales and vice versa.

    So the presence of bristles that “lie atop the scutes” strongly suggests that what we are looking at in these ornithischians are not protofeathers but some entirely different type of integument.

    No, it does not. Even accepting that feathers and scales cannot coexist, it is equally correct to conclude that ornithischians had true protofeathers and their scales were not homologous to those of non-ornithodirans.

    Along with these unique filamentous structures is their strange distribution. Monofilaments are found along the head, neck and thorax. The femur and humerus house the strange compound structures. The forearm, leg (tibia and fibula) and tail, meanwhile, contain scales.

    I’d say this distribution falsifies your earlier prediction: “Should the skin impression be found on the tail, pelvis, or thoracic region though, one should be able to reasonably infer that it was representative of the integument found across the body up to the head, ankle, and wrist regions.” In Kulindadromeus, no such inference can be made. Not only would skin impressions from the tail predict a totally different integument for the rest of the body than thorax impressions; neither region would allow us to infer the presence of the “bristly scales” along the femur and humerus.

    Still, if I were to predict placement, I would not be surprised if the Kulinda specimen wound up grouping closely with Tianyulong confuciusi, either as a heterodontosaurid, hypsilophodontid, or as something exclusive of the two.

    There is no need to “predict” where Kulindadromeus will wind up — it has already been included in what is probably the best (or at least the largest) data matrix available. While I agree that the low support values across the Godefroit et al. tree mean their hypothesis should be subjected to extensive testing in future analyses, eyeballing and observations such as “the Kulinda specimen shares many traits in common with Heterodontosaurus and Hypsilophodon , but also plenty of things that are different” (which means little more than that it’s an ornithischian) don’t constitute a valid test. A good test would be to download TNT and the original data matrix, constrain Kulindadromeus to be within heterodontosaurids, count the additional steps, and run the Templeton test to find out whether the difference is statistically significant. Without doing any of the above, your speculation that Kulindadromeus might be closely related to Tianyulong after all looks more like a slightly desperate attempt to make the hypothesis of multiple origins of ornithodiran filaments look less unparsimonious than it really is.

    Most dinosaurs are still scaly and it is still most parsimonious to assume that these structures evolved separately in ornithischians and advanced coelurosaurs.

    That, again, rests on the assumption that feathers and scales are mutually exclusive, which obviously isn’t true for the scuta and reticula in birds. How do we know the tuberculate scales visible in dinosaur skin impressions weren’t of the same origin and didn’t coexist with feathers that weren’t preserved due to the nature of the sediment?

    (BTW, where does your weird topology of ornithischians come from? According to a rather exhaustive overview at dinozaury.com, every single post-1985 analysis of ornithischians has recovered ceratopsians closer to ornithopods than to thyreophorans. According to Godefroit et al., Kulindadromeus is a basal member of a clade that doesn’t even exist on your tree.)

    Both groups appear to have ignored the extensive evolutionary developmental studies that have been performed on feather formation.

    That would be very hard to reconcile with your speculation that “Dhouailly’s contribution to the paper was mostly as a consultant for feather formation in birds”.

    In no case do feathers form from the distal portions of completed scales, which is what the Kulinda specimens seem to be showing. […] Dermal-epidermal recombination studies, in which the epidermis of one integument forming region is grafted to the dermis of another, has succeeded in achieving “feathery scales” (Rawles 1963). However these should not be looked at as evidence that structures could, or did exist in the past, as they involved heavy tissue recombination.

    The very papers you cite apparently contradict your assertions:

    “Histological examination of a hybrid structure from a Silkie embryo (Sawyer, ’70) showed that the scutate scale and feather epidermis were continuous within the structure. The base of the structure is a scutate scale, which developed from its placode and interplacode regions, but the feather filament is continuous with the scutate scale ridge and does not form from a feather placode. It develops directly into the epidermis of a feather filament while remaining continuous with the scutate scale epidermis (Sawyer, ’70).”

    —Sawyer & Knapp 2003: 66

    “Most of the time, the formation of the feathers, made of feather-type keratins (Dhouailly, unpublished data), do not render the scales unidentifiable, and are formed by continuous growth at the scale tip (Fig. 3B,D). This can occur in nature as shown by several cases of mutation in poultry (Somes, 1990).”

    —Dhouailly 2009: 592

    For comparison keep in mind that Squamata has evolved viviparity a conservative 100 times (Shine 2005)! In terms of complicated developmental processes, going from an external egg to an internally supported baby (including the evolution of a placenta) is way more difficult than sprouting a long, thin filament.

    I don’t see how the fact that some other feature has a highly homoplastic distribution somewhere else justifies your choice of a less parsimonious explanation for the distribution of filaments within Ornithodira (four or five origins instead of one).

  3. You have mentioned it repeatedly, but I don’t think you have ever provided a citation for it. The papers you cite suggest that feathers share many similarities with scales (including nonavian scales) in the placode stage and both evolved through a modification of the same ancestral developmental pathways (e.g., Sawyer et al. 2003; Dhouailly 2009). They also show there is an antagonistic relationship between feathers and all the various types of bird scales (scuta, scutella, reticula), which are formed by active feather suppression. However, the antagonism obviously doesn’t preclude their coexistence, since they do coexist on the body of every single living bird. The suppression is local, not global. Correct me if I’m wrong, but none of the many evo-devo papers you have cited in your articles on this topic actually supports your conclusion that the presence of feathers absolutely rules out the presence of scales and vice versa.

    I never meant to suggest that feathers and scales could not coexist on the same animal. As you rightly point out most birds do show the presence of feathers and scales on their body. However it is the regional differentiation that is important to keep in mind. Birds are often cited as examples of how dinosaurs could have been both feathery and scaly, but it is a false example as the scales are always localized to the tarsometatarsal region. They never stray from this spot. The various evo-devo studies that I have cited over the years continue to support the theory that scales have re-evolved in birds. The conventional thinking for dinosaur feather development is that it came on gradually with a few filaments here and scales everywhere else. The fact that birds do not show this during any point during development, and that not even in vitro manipulation can give a bird a scaly hide, all indicate that scales are a secondary derivation (and a localized one at that). This suggests that when feathers first formed they did so in toto. Again, the similarities between scales in reptiles and the scales and feathers in birds, occurs extremely early in development (the placode stage is unique to avian scales). The work by Dhouailly, Sawyer, Rawles and others suggest that the scale pathway was co-opted early in development to make feathers. This means that the association of feathers with scales is not as fluid as many paleontologists would believe.

    While I would not argue that the presence of feathers absolutely rules out the presence of scales, I would say that it strongly suggests it. That we have anomalous chickens and in vitro experiments showing that a “feathered scale” hybrid can happen, it obscures the point. The feathers are not growing along with the scale, they are growing in spite of the scale pathway’s inhibition of its formation. It’s an example of developmental crosstalk.

    Even accepting that feathers and scales cannot coexist, it is equally correct to conclude that ornithischians had true protofeathers and their scales were not homologous to those of non-ornithodirans.

    One could argue this, but it would be less parsimonious. Furthermore the burden of proof would be on the proponent to provide evidence that the scales on non-ornithodirans are secondarily derived. One would also need to show that the structures observed are homologous to protofeathers. That the scales on dinosaurs match up with the scales on other archosaurs, testudines, and squamates (those with non-imbricating scales) suggests that they are the same structures.

    It’s worth noting that bird scales do not really vary at all. We always have plate-like scutate scales on the dorsum of the tarsometatarsus and less plate-like scutellae along the sides and back of the tarsometatarsus. This could mean that scales are limited in scope in birds. It could also mean that the potential is there, but no bird lineage has “seen fit” to change things up.

    Regarding filament distribution in the Kulindadromeus:

    I’d say this distribution falsifies your earlier prediction: “Should the skin impression be found on the tail, pelvis, or thoracic region though, one should be able to reasonably infer that it was representative of the integument found across the body up to the head, ankle, and wrist regions.” In Kulindadromeus, no such inference can be made. Not only would skin impressions from the tail predict a totally different integument for the rest of the body than thorax impressions; neither region would allow us to infer the presence of the “bristly scales” along the femur and humerus.

    Yes and no. If this was a coelurosaur I would argue that these impressions would falsify my prediction. We have a good handle on how and where feathers develop along the skin in birds, and we can use this as a “roadmap” for dinosaurs that were close to birds. Kulindadromeus lies pretty darned far from birds. That it appears to challenge the “rule” of integument distribution is interesting, but also highly suggestive of these structures not being homologous to protofeathers.

    There is no need to “predict” where Kulindadromeus will wind up — it has already been included in what is probably the best (or at least the largest) data matrix available. While I agree that the low support values across the Godefroit et al. tree mean their hypothesis should be subjected to extensive testing in future analyses, eyeballing and observations such as “the Kulinda specimen shares many traits in common with Heterodontosaurus and Hypsilophodon , but also plenty of things that are different” (which means little more than that it’s an ornithischian) don’t constitute a valid test. A good test would be to download TNT and the original data matrix, constrain Kulindadromeus to be within heterodontosaurids, count the additional steps, and run the Templeton test to find out whether the difference is statistically significant. Without doing any of the above, your speculation that Kulindadromeus might be closely related to Tianyulong after all looks more like a slightly desperate attempt to make the hypothesis of multiple origins of ornithodiran filaments look less unparsimonious than it really is.

    I don’t see any harm in making a prediction of future placement. I also predicted that Dilong paradoxus would wind up falling closer to Maniraptora than Tyrannosauroidea in the future. Though I did not perform any phylogenetic test of this myself, both Turner et al. 2007 and Lee & Worthy 2011 did find exactly that. It doesn’t mean it’s right, but it does suggest that the characters holding Dilong in Tyrannosauroidea are pretty weak (the low support value for that node also suggests this). It should also be pointed out that Zheng et al’s phylogenetic analysis of Tianyulong did not provide any statistical information for any of its nodes either, so calling Tianyulong a heterodontosaurid may be premature.

    That said, I will take your point into consideration. However, even if I come up with a result that is a few steps shorter it won’t much matter without statistical support for the topology (which is a character problem). It would be interesting to see what a Bayesian analysis would produce.

    Regarding making scales plesiomorphic for Dinosauria:

    That, again, rests on the assumption that feathers and scales are mutually exclusive, which obviously isn’t true for the scuta and reticula in birds.

    True, given the assumptions I laid out in my posts, it would be most parsimonious to assume scales as the plesiomorphic state for Dinosauria. However, keep in mind that Barrett’s study also found this to be true. I don’t think he demanded mutual exclusivity for his analyses. It’s important not to confuse the superficially close relationship of feathers and scales in birds, with their potential in dinosaurs. For instance, reticula and feathers should coexist just fine since reticula are most likely aborted feathers. Whether or not scutate scales are secondarily derived in dinosaurs can be argued, and potentially even tested (e.g., the scutate scales preserved in Concavenator corvatus could be compared [using SEM] to the scutate scales preserved in NGMC-91). The diversity of dinosaur scales compared to bird scales would suggest that they were reptilian, rather than avian, in structure (assuming that avian scales are inherently limited).

    How do we know the tuberculate scales visible in dinosaur skin impressions weren’t of the same origin and didn’t coexist with feathers that weren’t preserved due to the nature of the sediment?

    My problem with the taphonomic argument is that we don’t see the problem that people worry about. That is to say we don’t find scale impressions to the exclusion of filament impressions in animals that had both. Nor do we see the reverse. We either find integument impressions or we don’t. Pelacanimimus polydon and Scipionyx samniticus are both good examples of how taphonomy can be selective in its preservation, yet still treat scales and feathers the same. A more classic example is Archaeopteryx lithographica and Compsognathus longipes. The former preserves feathers whereas the latter preserves nothing. No scales, no feathers, no soft tissue at all. More recent analyses on Archaeopteryx (Foth et al. 2014) have found evidence for feathered hind limbs too, indicating that the lack of scale preservation along the feet is due to scales not being there in the first place. Given that scales and feathers are both epidermal structures it makes sense that whatever preservation that favours one should favour the other as well.

    (BTW, where does your weird topology of ornithischians come from? According to a rather exhaustive overview at dinozaury.com, every single post-1985 analysis of ornithischians has recovered ceratopsians closer to ornithopods than to thyreophorans. According to Godefroit et al., Kulindadromeus is a basal member of a clade that doesn’t even exist on your tree.)

    When I first compiled that tree I was using topologies presented on Wikipedia. So either I copied over false information from Wikipedia at the time (possible) or I transposed something when I was originally making the image (also possible). I’ll fix the error.

    Regarding authors ignoring the evo-devo literature:

    That would be very hard to reconcile with your speculation that “Dhouailly’s contribution to the paper was mostly as a consultant for feather formation in birds”.

    Consultant does not necessarily mean that all suggested changes were used. Remember, the authors were more cautious of their interpretations in the supplement than in the main paper. I suspect a degree of publication “sexiness” may have been at play.

    Regarding “feathery scales”:

    The very papers you cite apparently contradict your assertions:
    “Histological examination of a hybrid structure from a Silkie embryo (Sawyer, ’70) showed that the scutate scale and feather epidermis were continuous within the structure. The base of the structure is a scutate scale, which developed from its placode and interplacode regions, but the feather filament is continuous with the scutate scale ridge and does not form from a feather placode. It develops directly into the epidermis of a feather filament while remaining continuous with the scutate scale epidermis (Sawyer, ’70).”
    —Sawyer & Knapp 2003: 66

    “Most of the time, the formation of the feathers, made of feather-type keratins (Dhouailly, unpublished data), do not render the scales unidentifiable, and are formed by continuous growth at the scale tip (Fig. 3B,D). This can occur in nature as shown by several cases of mutation in poultry (Somes, 1990).”
    —Dhouailly 2009: 592

    Be wary of data obtained from artificially bred animals. Silkies also routinely produce five toes, with the first and second toe being webbed. Just as with the in vitro experiments these data are showing us how the pathways work, which can help explain how these structures evolved. Hence why Sawyer and Knapp also say:

    “These data suggest that scutate scale development is dependent on gene expression occurring in the dermis around day 8 of development and may involve an inhibition of feather formation.”””pg: 62

    Similarly, Dhouailly 2009 mentions:

    “Whereas the scuta-feather metaplasia happens in nature through mutation, and is very easily obtained in various experimental conditions, the reverse, the feather-scuta metaplasia, has never occurred naturally or been obtained experimentally.”—pg: 594

    “The overlapping avian scuta and scutella which cover the tarsometatarsi and dorsal digits of the feet in some bird lineages thus appear to be secondarily derived from feathers”””pg: 602

    And since we are on the subject, in his review of feather evolution, Sawyer 2005 wrote:

    “From these observations [chicken mutants] and studies using experimental epidermal— dermal tissue recombinations (Rawles, 1963; Sengel and Abbott, 1963; Goetinck and Abbott, 1963), the concept emerged that scale formation on the legs and feet of today’s birds first requires a repression of feather development (Brotman, 1977a—c). An evolutionary extension of this concept suggests that the legs and feet of the ancestor of birds were feathered and that the scutate scales of today’s birds are derived from feathers (Sawyer and Knapp, 2003a). The recent discovery of Microraptor, a dinosaur with feathered fore- and hindlimbs (Xu et al. 2003), adds support to this view”

    I don’t see how the fact that some other feature has a highly homoplastic distribution somewhere else justifies your choice of a less parsimonious explanation for the distribution of filaments within Ornithodira (four or five origins instead of one).

    As I mentioned above, it’s the concept that a structure as “complicated as a feather” could have evolved more than once, that doesn’t sit well with some folks in paleo. Viviparity in squamates stands as an elegant example of a highly complicated structure evolving on numerous occasions. One needs to also keep in mind that we are not just talking about the evolution of feathers in Dinosauria, or Ornithodira. We are also talking about the re-evolution of scales in multiple lineages (up to seven times if we assume that filaments appeared once). Current evo-devo studies indicate that scales and feathers are not “hot-swappable”. You can’t just turn off feather formation and wind up with scales (if we did, then these things would not exist). Feathers need to be actively inhibited in order for scales to form in the only region of the bird body that scales ever form.

    I would also like to point out that the homology of these filaments are still only assumed. The only tests of homology that have been done on dinosaur filaments are the protofeathers of maniraptorans (Xu et al. 2010). Even then it was more a refutation of Lingham-Soliar’s work than an actual primary test of homology. The only other test was that for secondary homology of filaments across Ornithodira (Barrett 2013). Whereas Xu et al. offered some support for primary homology of advanced coelurosaur filaments and bird feathers, the other””broader test””did not find support for filaments being ancestral to Dinosauria, much less Ornithodira.

    It is misleading to assume homology without first testing it. This has lead to all kinds of problems in the muscle literature, where researchers originally were far too willing to give similar muscles the exact same names. This will likely lead to many problems in the feather literature too if folks aren’t careful. Sawyer et al. (2003) had urged caution with this already. I’m largely just echoing their point.

    Extra References

    Lee, M. S. Y., Worthy, T.H.. 2011. Likelihood Reinstates Archaeopteryx as a Primitive Bird. Biology Letters. Vol. 8(2):299—303.

    Turner, A.H., Pol, D., Clarke, J.A., Erickson, G.M., Norell, M.A. 2007. A Basal Dromaeosaurid and Size Evolution Preceding Avian Flight. Science. Vol.317:1378—1381.

    Sawyer, R.H., Rogers, L., Washington, L., Glenn, T.C., Kapp, L.W. 2005. Evolutionary Origin of the Feather Epidermis. Dev. Dyn. Vol. 232:256—267.

    Sawyer, R.H., Washington, L.D., Salvatore, B.A., Glenn, T.C., Knapp, L.W. 2003b. Origin of Archosaurian Integumentary Appendages: The Bristles of the Wild Turkey Beard Express Feather-Type β-Keratins. J. Exp. Zool. (Mol. Dev. Evol.) Vol. 297B:27—34.

    Xu,X., Zheng, X., You, H. 2010. Exceptional Dinosaur Fossils Show Ontogenetic Development of Early Feathers. Nature. Vol. 464 (Supplement):1—12.

  4. However it is the regional differentiation that is important to keep in mind. Birds are often cited as examples of how dinosaurs could have been both feathery and scaly, but it is a false example as the scales are always localized to the tarsometatarsal region.

    I’m not sure why that would be a false example. Birds are obviously highly conservative in their integument disribution, but can we safely extrapolate from them to all nonavian dinosaurs? It’s true that the tarsometatarsus (plus the distal part of the tibiotarsus) is the only body region in living birds that shows “integumentary plasticity”, i.e., the ability to form more than one type of integument, but is it really so inconceivable that this “zone of plasticity” might have been larger in nonavian dinosaurs — perhaps even extending over most of the body?

    The conventional thinking for dinosaur feather development is that it came on gradually with a few filaments here and scales everywhere else.

    Do you have a citation for that? As far as I know, insulation continues to be one of the most popular hypotheses about the original function of feathers (it’s cited favorably by Godefroit et al., for example), and if the first feathers were to have had any insulatory effect at all, they would have had to cover a large part of the body. Xu & Guo (2009) regard the sparse distribution of filaments in Psittacosaurus and Tianyulong as representative of the ancestral condition (which is weird — what about pterosaurs?), but they do not comment on whether the rest of the body was covered by scales or naked skin.

    The fact that birds do not show this during any point during development, and that not even in vitro manipulation can give a bird a scaly hide, all indicate that scales are a secondary derivation (and a localized one at that). This suggests that when feathers first formed they did so in toto.

    I don’t see how the second sentence follows from the first. Whatever we learn from evo-devo research about the relationship between feathers and avian scales (which, as we agree, probably represent a secondary derivation) doesn’t tell us much about the possibility of coexistence between feathers and primary, plesiomorphic archosaur scales. Your second point about feathers evolving through a modification to the developmental pathway of scales is more relevant in this respect, but I don’t see how that is supposed to prevent them from coexisting, either. What if the modification was local at first? Is there any good reason to think it couldn’t have been local and must have impacted the whole surface of the body from the very beginning? Remember, the dinosaurs with the most extensive feathering (such as Anchiornis with pennaceous feathers attached to the pedal phalanges — Hu et al. 2009) are relatively derived.

    Again, the similarities between scales in reptiles and the scales and feathers in birds, occurs extremely early in development (the placode stage is unique to avian scales).

    Sorry, bad wording on my part — I didn’t mean to imply that nonavian scales have the placode stage as well.

    The work by Dhouailly, Sawyer, Rawles and others suggest that the scale pathway was co-opted early in development to make feathers. This means that the association of feathers with scales is not as fluid as many paleontologists would believe.

    Again, I’m not convinced that your conclusions follow from the papers you cite. (In fact, you rather consistently cite Dhouailly’s research in indirect support of statements that her own paper now contradicts; more on that below.) For example, the way Sawyer et al. (2003) phrased their hypothesis hardly suggests it is impossible for feathers and scales to coexist:

    “The similarities in the epidermal cell populations that make up the feather and the embryonic layers of the developing scutate scale (Fig. 1), as well as the embryonic epidermis of the alligator (Alibardi and Thompson, 2001), suggest that the epidermal appendages of crocodilians and birds evolved through the developmental modification of epidermal cell populations, which characterized the integument of ancestral archosaurians. Most likely, these epidermal populations had already evolved the ability to express different members of the α and/or β keratin gene families. Developmental changes in the spatial and temporal expression of genes regulating pattern formation and signaling systems would have created the integumentary structures seen in living archosaurians.”

    —Sawyer et al. 2003: 20

    That we have anomalous chickens and in vitro experiments showing that a “feathered scale” hybrid can happen, it obscures the point. The feathers are not growing along with the scale, they are growing in spite of the scale pathway’s inhibition of its formation. It’s an example of developmental crosstalk.

    I’d say it is the specific developmental mechanisms which produce feather-bearing scales that are irrelevant here and obscure the point. What is important is that feathered scales do occur in nature, and they don’t even appear to be particularly rare (another paper that you cite in the article — Rawles 1963 — says that much in the introduction).

    One could argue this, but it would be less parsimonious.

    Perhaps it would, but note that adding Yutyrannus and Sciurumimus to your tree and rearranging the ornithischians according to the modern consensus leaves both scenarios equally parsimonious.

    Furthermore the burden of proof would be on the proponent to provide evidence that the scales on non-ornithodirans are secondarily derived.

    Surely you mean the scales on ornithodirans?

    One would also need to show that the structures observed are homologous to protofeathers. That the scales on dinosaurs match up with the scales on other archosaurs, testudines, and squamates (those with non-imbricating scales) suggests that they are the same structures.

    This seems like a double standard to me — when it comes to the homology of ornithodiran and non-ornithodiran scales, you do not employ the same rigorous criteria you are applying to feathers. If a high degree of overall similarity is enough to establish the homology of scales, why is it not enough for feather-like filaments?

    If this was a coelurosaur I would argue that these impressions would falsify my prediction. We have a good handle on how and where feathers develop along the skin in birds, and we can use this as a “roadmap” for dinosaurs that were close to birds. Kulindadromeus lies pretty darned far from birds.

    I don’t see any reason for restricting the prediction to coelurosaurs. Do you have any objective criterion for determining which dinosaurs are close enough to birds that we can use bird feather development as a viable roadmap for them? But never mind: if only coelurosaurs count, what about Juravenator, which has an even more scattered distribution of feathers and scales than Kulindadromeus (with both co-occuring on the tail, for example)?

    It should also be pointed out that Zheng et al’s phylogenetic analysis of Tianyulong did not provide any statistical information for any of its nodes either, so calling Tianyulong a heterodontosaurid may be premature.

    Weakly supported, yes, but not premature — the placement in Heterodontosauridae has been repeatedly shown to be the best interpretation of available data.

    That said, I will take your point into consideration. However, even if I come up with a result that is a few steps shorter it won’t much matter without statistical support for the topology (which is a character problem).

    I would be extremely surprised if you managed to find a shorter tree — that would mean that even with all the powerful heuristics available in TNT, Godefroit et al.’s analysis failed to recover the best tree(s) for a relatively tiny dataset. Or did you mean re-running the analysis after some changes to the original coding? (I agree with your point about statistical support, of course.)

    It would be interesting to see what a Bayesian analysis would produce.

    Agreed, but given that parsimony and Bayesian inference almost always find the same tree when applied to morphological data, it’s reasonable to assume the Bayesian topology wouldn’t differ much from the current consensus.

    However, keep in mind that Barrett’s study also found this to be true. I don’t think he demanded mutual exclusivity for his analyses.

    I think he did, because the results you get when you relax that assumption (for example, by inferring the ancestral states separately for scales and feathers) are totally different from those that Barrett and Evans presented at the SVP meeting:

    http://theropoda.blogspot.cz/2014/07/un-approccio-probabilistico.html

    My problem with the taphonomic argument is that we don’t see the problem that people worry about. That is to say we don’t find scale impressions to the exclusion of filament impressions in animals that had both.

    I believe there is a number of scale impressions from tyrannosaurids, which are predicted to have had feathers as well based on Type I phylogenetic bracketing (anchored on maniraptoriforms and Yutyrannus, which has been recovered as a derived tyrannosauroid with two different data matrices and considerable statistical support from both parsimony and Bayesian analyses — Xu et al. 2012; Lee et al. 2014).

    Given that scales and feathers are both epidermal structures it makes sense that whatever preservation that favours one should favour the other as well.

    That’s not how taphonomy works. In particular, feathers are much more delicate than scales and might not be preserved in coarser-grained sediments that are still fine enough for the preservation of scale impressions.

    Consultant does not necessarily mean that all suggested changes were used. Remember, the authors were more cautious of their interpretations in the supplement than in the main paper. I suspect a degree of publication “sexiness” may have been at play.

    While this is all true, I can’t help thinking it’s an increasingly strained attempt to explain away why Dhouailly’s own paper disagrees with the conclusions you draw from her earlier research. We don’t know if she personally examined the material from Kulinda (which is a shame — in many other journals, listing individual authors’ contributions is a common practice), but it’s quite unbelievable that she would agree to co-author a paper that fails to take her most basic suggestions into account and promotes a hypothesis she doesn’t agree with. Therefore, I have no reason to doubt that statements such as “protofeather-like structures were probably widespread in Dinosauria, possibly even in the earliest members of the clade” (Godefroit et al. 2014: 455) and “it is possible that the extensively scaled distal hindlimbs in Kulindadromeus might be explained by similar local and partial inhibition in the development of featherlike structures” (ibid) represent Dhouailly’s opinion as well, and that she considers them to be compatible with her own evo-devo research.

    Incidentally, I don’t share the impression that the authors were more cautious in the supplementary materials. I think the journal’s severe space limits made them go with the interpretation they considered most likely in the main text, while the supplement offered a plenty of space for a more detailed description.

    Be wary of data obtained from artificially bred animals. Silkies also routinely produce five toes, with the first and second toe being webbed.

    Wow — I didn’t know that, thanks for the cool information! However, even if that somehow disqualified their feather-bearing scales from further consideration, Silkies are far from the only birds with this kind of structures, and I don’t see how your quotes from Sawyer & Knapp (2003) and Dhouailly (2009) could change that. Rawles (1963) lists several less aberrant breeds as exhibiting the same condition (“Silver Campine, White Leghorn and others” — p. 49) and emphasizes that feather-bearing scales “were clearly identified by many of the earlier workers” (ibid). Unfortunately, I don’t have access to the number of ancient papers (mostly written in German) she cites in support of this statement.

    As I mentioned above, it’s the concept that a structure as “complicated as a feather” could have evolved more than once, that doesn’t sit well with some folks in paleo. Viviparity in squamates stands as an elegant example of a highly complicated structure evolving on numerous occasions.

    I don’t think that anyone actually rules out the possibility of things like this happening — the “folks in paleo” simply don’t want to postulate them if it isn’t necessary. While your series of well-researched, well-referenced articles represents the best attempt to show it is in fact necessary that I’m aware of (and I cite it as such on my blog), it still fails to convince me to accept the less parsimonious scenario.

    It is misleading to assume homology without first testing it.

    I disagree. I view the (repeatedly noted) overall similarity between the filaments in ornithischians and tetanurines as sufficient for applying Hennig’s auxiliary principle to them: they are to be considered homologous until proven otherwise. Perhaps Barrett and Evans’s unpublished ancestral state analysis presents such a proof to the contrary, but then again, Andrea Cau’s unpublished analysis that I linked to above does most certainly not.

    Refs:

    Hu D-Y, Hou L-H, Zhang L-J, Xu X 2009 A pre-Archaeopteryx troodontid theropod from China with long feathers on the metatarsus. Nature 461(7264): 640—3

    Lee MSY, Cau A, Naish D, Dyke GJ 2014 Sustained miniaturization and anatomical innovation in the dinosaurian ancestors of birds. Science 345(6196): 562—6

    Xu X, Guo Y 2009 The origin and early evolution of feathers: insights from recent paleontological and neontological data. Vert PalAs 47(4): 311—29

    Xu X, Wang K-B, Zhang K, Ma Q-Y, Xing L-D, Sullivan C, Hu D-Y, Cheng S-Q, Wang S 2012 A gigantic feathered dinosaur from the Lower Cretaceous of China. Nature 484(7392): 92—5

  5. I’m not sure why that would be a false example. Birds are obviously highly conservative in their integument disribution, but can we safely extrapolate from them to all nonavian dinosaurs? It’s true that the tarsometatarsus (plus the distal part of the tibiotarsus) is the only body region in living birds that shows “integumentary plasticity”, i.e., the ability to form more than one type of integument, but is it really so inconceivable that this “zone of plasticity” might have been larger in nonavian dinosaurs — perhaps even extending over most of the body?

    Though it is possible that various dinosaur lineages “amped up” a re-evolution of scales such that a more dynamic association of feathers and scales was possible, it is not something that is borne out by developmental studies. Remember Rawles, as well as Dhouailly (1973), in their recombination studies, found that only the dermis of birds has the capacity to produce scales. Everywhere else on the body that was tested, would only produce feathers (except for the beak, which always produced a beak regardless of what epidermis was placed on it). This is strongly suggestive that the lineage leading to birds only ever re-evolved scales on the foot region. The presence of scales on the foot of NGMC-91, but protofeathers everywhere else, further supports this scenario. Since scales rarely preserve in the fossil record, and scale development doesn’t preserve at all, we are left with the extant phylogenetic bracket to determine scale type in dinosaurs. However the distribution of scales on extant archosaurs doesn’t help us much here either. Developmental work has shown that the scales of birds develop differently than the scales of crocodylians (see Alibardi and Thompson 2001 for the croc stuff). Since dinosaurs were in the middle, they could have had crocodylian-like scales, avian-like scales, or (more likely) were in the spectrum that lead from one to the other. If we assume that filaments homologous to primordial feathers were ancestral to Ornithodira and that multiple dinosaur species re-evolved scales en masse using a developmental mechanism that appears to be nonexistent in birds, then we wind up making a level 3′ inference for the ancestral character state (a structure present in the extinct taxon that doesn’t leave an osteological correlate and is not found in the extant relatives). It basically just special pleading at that point.

    Do you have a citation for that? As far as I know, insulation continues to be one of the most popular hypotheses about the original function of feathers (it’s cited favorably by Godefroit et al., for example), and if the first feathers were to have had any insulatory effect at all, they would have had to cover a large part of the body. Xu & Guo (2009) regard the sparse distribution of filaments in Psittacosaurus and Tianyulong as representative of the ancestral condition (which is weird — what about pterosaurs?), but they do not comment on whether the rest of the body was covered by scales or naked skin.

    Dyck 1985 gives a good summary of all the previous theories for the evolution of feathers. Back when feathers were viewed as primarily aerodynamic structures, many proponents argued for a start along the legs and arms first. This was seen a lot in the old Discovery dino documentaries and can still be seen in various paleo-art reconstructions today (the one’s that half-ass the feathering). This concept of slow overtaking of the integument also works if filaments served a display or anti-predatory role. It becomes less viable for the thermoregulatory role, which happens to be the popular hypothesis at the moment. Surprisingly, despite the popularity of the insulatory model, I still hear proponents argue for the old concept of feathers being more sparsely distributed ancestrally (e.g, the ever popular “feathered mohawk” seen in the first Tianyulong reconstruction, along with many other dinosaur reconstructions. This concept pops up from time to time on the DML too). As you point out, Xu and Guo also argue for an ancestrally sparse distribution to filaments, though in their case they don’t argue for an initial thermoregulatory function to filaments.

    I don’t see how the second sentence follows from the first. Whatever we learn from evo-devo research about the relationship between feathers and avian scales (which, as we agree, probably represent a secondary derivation) doesn’t tell us much about the possibility of coexistence between feathers and primary, plesiomorphic archosaur scales. Your second point about feathers evolving through a modification to the developmental pathway of scales is more relevant in this respect, but I don’t see how that is supposed to prevent them from coexisting, either. What if the modification was local at first? Is there any good reason to think it couldn’t have been local and must have impacted the whole surface of the body from the very beginning?

    The fact that feathers are an apparently ancestral structure for all the skin on birds is what strongly suggests that there was a total replacement of integument. If there was only a partial replacement over time””as has been suggested for feather evolution before””then why would birds need to re-evolve scales again? Why not just “turn off” feather development and “turn on” scale development? Developmental data strongly suggest that birds don’t do it this way because they can’t. Avian-scale epidermis also doesn’t understand the scale-forming signal of lizard dermis (Dhouailly 1975). This suggests birds lost the original scale-forming instructions altogether. If those instructions were “hijacked” to form feathers instead, it would make sense why bird skin can’t form reptile scales.

    Remember, the dinosaurs with the most extensive feathering (such as Anchiornis with pennaceous feathers attached to the pedal phalanges — Hu et al. 2009) are relatively derived.

    Well Microraptor is (or was) considered more basal than Sinornithosaurus (though they might be considered sister-taxa now). Anchiornis used to be more basal than that, though I guess it is considered a avialan this week. That Confuciusornis also shows extensive feathering and is definitely more apical than dromaeosaurs suggests that birds were doing the feathers everywhere thing for quite a long time before they started to dial things back. Dromaeosaurs seemed to have dialed things back independently at an earlier stage (or NGMC 91 and””by association””Sinornithosaurus were unique).

    Again, I’m not convinced that your conclusions follow from the papers you cite. (In fact, you rather consistently cite Dhouailly’s research in indirect support of statements that her own paper now contradicts; more on that below.) For example, the way Sawyer et al. (2003) phrased their hypothesis hardly suggests it is impossible for feathers and scales to coexist:

    “The similarities in the epidermal cell populations that make up the feather and the embryonic layers of the developing scutate scale (Fig. 1), as well as the embryonic epidermis of the alligator (Alibardi and Thompson, 2001), suggest that the epidermal appendages of crocodilians and birds evolved through the developmental modification of epidermal cell populations, which characterized the integument of ancestral archosaurians. Most likely, these epidermal populations had already evolved the ability to express different members of the α and/or β keratin gene families. Developmental changes in the spatial and temporal expression of genes regulating pattern formation and signaling systems would have created the integumentary structures seen in living archosaurians.”

    —Sawyer et al. 2003: 20

    Sawyer et al. were talking about the specific embryonic epidermal layers present in crocodylians and birds during integument development. Both structures produce feather β-keratins. In birds these epidermal layers form the barb ridge and embryonic feather filament, whereas in the avian-scales and crocodylian scales, the embryonic epidermis appears to be largely a protective covering that is sloughed prior to hatching (Sawyer et al. 2005). The authors argue that these layers are what gave rise to feathers. Their presence in crocs suggests that these layers were present in the MRCA of crocs and birds and it is these layers that are responsible for the various types of filaments seen in archosaurs. When the embryonic layers are augmented, scale development is replaced by feathers. When scales are allowed to develop, the feather-forming layers get dumped. Only in strange developmental scenarios (often caused via in vitro tampering) can the crosstalk happen and one get “feathery scales” in birds. I would argue that the best test for feather-scale non-antagonistic association would be to see if one could get protofeathers to sprout from the embryonic epidermal layers in developing crocodylians (in this case, a filament with a barb ridge). If they can form without interfering with scale formation then that would be stronger evidence for a scale-feather association.

    I’d say it is the specific developmental mechanisms which produce feather-bearing scales that are irrelevant here and obscure the point. What is important is that feathered scales do occur in nature, and they don’t even appear to be particularly rare (another paper that you cite in the article — Rawles 1963 — says that much in the introduction).

    Keep in mind that Rawles discussed this as not being rare in the chicken breeds that she cited. They are also found in artificially bred pigeons. Nowhere does she mention this happening in wild birds. As I mentioned previously, artificially bred animals tend to show wonky developmental stuff anyway. It gives us ideas of what the mechanisms were but might not be as useful for assessing prior phenotypes. The only birds that I can think of that routinely have a feather-scale mix on their feet are owls. Their apparent easy association is extremely interesting and I would love to know if the filaments coming between the scales actually are feathers or if they are unique filaments akin to the turkey beard. That reminds me, I have access to a frozen barn owl. I should see if I can get samples from its foot.

    Perhaps it would, but note that adding Yutyrannus and Sciurumimus to your tree and rearranging the ornithischians according to the modern consensus leaves both scenarios equally parsimonious.

    I’m skeptical of Sciurumimus both in terms of preservation and where it sits among dinosaurs. Yutyrannus could add one more step, but given its relationship to other coelurosaurs (it at least seems to stay a coelurosaur, Andrea Cau’s analysis not withstanding) one could also just push the filament node down a bit from Maniraptora. If Yutyrannus does stay as a tyrannosauroid then there still would not need to be an extra step as the filament node would just engulf practically all Coelurosauria. People have suggested that before. It would make the scales of tyrannosaurids, secondarily derived. That would count as evidence for dinosaurs re-evolving scales all over vs. just at the feet. Of course I’ve also heard people argue that tyrannosaurs just had naked skin rather than scaly skin. I’m less sold on this argument. Wyrex looks pretty scaly to me.

    Surely you mean the scales on ornithodirans?

    Heh, yep. My mistake there.

    This seems like a double standard to me — when it comes to the homology of ornithodiran and non-ornithodiran scales, you do not employ the same rigorous criteria you are applying to feathers. If a high degree of overall similarity is enough to establish the homology of scales, why is it not enough for feather-like filaments?

    For one there is not a high degree of structural similarity between the structures seen in some dinosaurs and bird feathers. Pterosaurs, Tianyulong, Kulindadromeus and that Psittacosaurus specimen all feature structures that are similar to feathers only at the most basic level. This is the same level that turkey beards, mammal hair and insect setae are similar to feathers. It’s an extremely ambiguous stage. Only once inside Maniraptora do these filaments start to show a high degree of similarity to bird feathers. Hence why I have no qualms with calling Maniraptoran filaments protofeathers, but urge caution when talking about the structures on Sinosauropteryx pterosaurs and the ornithischians mentioned. As I said earlier in this thread, Sawyer et al. (2003b) were the first to warn against assuming all filaments found in dinosaurs should be called feathers. I might be more vocal about it, but this is hardly unique to me.

    All that said I am all in favour of looking at the ultrastructure of fossilized scales to see if we can determine what their histological makeup was. We are starting to see a call from this in the paleo community (e.g., Bell et al. 2014). Keep in mind that when I said the variably-shaped scales on dinosaurs matched the type of variation we see in extant reptiles as opposed to extant birds, I did couch it in the assumption that bird skin has a limited ability to make scales. Should that prove untrue, then that would leave everything more ambiguous. I think the best comparison would be to look at fossilized scales from a turtle or crocodyliform and see how they compare to the scales on various dinosaurs. I would predict that if a difference could be determined it would probably pop up around Dromaeosauridae. Of course the biggest hurdle would be finding actual fossilized scales vs. just their impressions.

    I don’t see any reason for restricting the prediction to coelurosaurs. Do you have any objective criterion for determining which dinosaurs are close enough to birds that we can use bird feather development as a viable roadmap for them?

    Coelurosauria is where most of the shared similarities between birds and dinosaurs start to show up, including the formation of protofeathers. Since we are interested in when bird feather development first started, using the animals that first show signs of unambiguous feathers would seem like a logical start.

    But never mind: if only coelurosaurs count, what about Juravenator, which has an even more scattered distribution of feathers and scales than Kulindadromeus (with both co-occuring on the tail, for example)?

    I’ve mentioned my misgivings on Juravenator before. Chiappe and Gohlich argued that they found “extremely thin filaments at the edge of the preserved soft-tissue.” These “filaments” were so small that even under high magnification they could only be partially observed. Their figure 25 is supposed to show them, but all we get is a blurry image. If these are protofeathers, then they are the thinnest protofeathers yet found. I suspect they are most likely artifacts of either preservation or preparation.

    Weakly supported, yes, but not premature — the placement in Heterodontosauridae has been repeatedly shown to be the best interpretation of available data.

    Repeated placement could just as easily mean that something in the data is drawing Tianyulong towards that node (such as vague characters that are plesiomorphic to many members of the group, which is the problem Kulindadromeus suffers from). Indeed, weak bootstrap and Bremer support values both suggest that there are a couple of characters that are overly-influencing the placement.

    Regarding Barrett’s study assuming mutual exclusivity of feathers and scales:

    I think he did, because the results you get when you relax that assumption (for example, by inferring the ancestral states separately for scales and feathers) are totally different from those that Barrett and Evans presented at the SVP meeting:
    http://theropoda.blogspot.cz/2014/07/un-approccio-probabilistico.html

    It would like to know more details on Andrea’s methodology for his study. For instance what would happen if filaments, feathers and scales were all analyzed at once? It’s interesting that his results indicate that filaments may have been present in crocodylomorphs as well. I know that Barrett and Evans looked at scale preservation along with body distribution of scales and filaments. If they were only doing a binary comparison like Andrea’s then there should be no reason to care where the scales and filaments were found, only that they were present. I contacted Barrett about this and he said that the filament homology paper is currently in review (it went into review a few days before Kulindadromeus dropped) so, barring any significant problems, it should be out soon.

    I believe there is a number of scale impressions from tyrannosaurids, which are predicted to have had feathers as well based on Type I phylogenetic bracketing (anchored on maniraptoriforms and Yutyrannus, which has been recovered as a derived tyrannosauroid with two different data matrices and considerable statistical support from both parsimony and Bayesian analyses — Xu et al. 2012; Lee et al. 2014).

    The descriptive paper on Yutyrannus did not have strong statistical support at all and was criticized for its matrix choice. I had not looked that closely at the Lee et al. tree before. I think this might represent the first time that Tyrannosauroidea has been recovered with any kind of strong statistical support before (albeit a likely overestimated support due to the vagaries of posterior probability scoring). That’s pretty neat.

    Regarding filaments vs. scales preservation.

    That’s not how taphonomy works. In particular, feathers are much more delicate than scales and might not be preserved in coarser-grained sediments that are still fine enough for the preservation of scale impressions.

    As I mentioned above, that’s the argument that is typically put forth. However it is not what we actually ever see. Either feathers and scales are both preserved, or they both aren’t. Also the coarse-grained sediment argument seems less likely now that we have found filament preservation in sandstone (Zelenitsky et al. 2012).

    While this is all true, I can’t help thinking it’s an increasingly strained attempt to explain away why Dhouailly’s own paper disagrees with the conclusions you draw from her earlier research. We don’t know if she personally examined the material from Kulinda (which is a shame — in many other journals, listing individual authors’ contributions is a common practice), but it’s quite unbelievable that she would agree to co-author a paper that fails to take her most basic suggestions into account and promotes a hypothesis she doesn’t agree with. Therefore, I have no reason to doubt that statements such as “protofeather-like structures were probably widespread in Dinosauria, possibly even in the earliest members of the clade” (Godefroit et al. 2014: 455) and “it is possible that the extensively scaled distal hindlimbs in Kulindadromeus might be explained by similar local and partial inhibition in the development of featherlike structures” (ibid) represent Dhouailly’s opinion as well, and that she considers them to be compatible with her own evo-devo research.

    I contacted Dhouailly about this asking if she had a chance to look at the specimen itself. She does seem okay with calling these structures feathers, but stated that this finding does conflict with current interpretations of feather evolution (including her 2009 study). However she felt more comfortable rewriting what we currently know about feather evo-devo rather than suggesting that these were convergent structures. For that I’m not sure why. One thing to keep in mind is that she doesn’t do work on dinosaurs normally, so the large gulf between protofeathered coelurosaurs and filamented ornithischians (with lots of scaly guys in between) was probably not something she was aware of. I never got a response regarding a first-hand look at the specimen. However she did say that she is working on a follow-up to her 2009 article on feather development. This one is going to incorporate more fossil material in it (potentially including pterosaurs) So we’ll see what comes of that.

    I don’t think that anyone actually rules out the possibility of things like this happening — the “folks in paleo” simply don’t want to postulate them if it isn’t necessary. While your series of well-researched, well-referenced articles represents the best attempt to show it is in fact necessary that I’m aware of (and I cite it as such on my blog), it still fails to convince me to accept the less parsimonious scenario.

    That’s all fine and good. I keep the argument out there because I do think this is an important problem that needs to be addressed. I also don’t accept the less parsimonious scenario. We just disagree on what that scenario is.

    Extra references

    Alibardi, L, Thompson, M.B. 2001. Fine Structure of the Developing Epidermis in the Embryo of the American Alligator (Alligator mississippiensis, Crocodilia, Reptilia). J. Anat. Vol. 198:265—282.

    Bell, P.R., Fanti, F., Currie, P.J., Arbour, V.M. 2014. A Mummified Duck-Billed Dinosaur with a Soft-Tissue Cock’s Comb. Current. Biol. Vol. 24:1—6.

    Dhouailly, D. 1975. Formation of Cutaneous Appendages in Dermo-Epidermal Recombinations between Reptiles, Birds and Mammals. Wilhelm Roux’s Archives. Vol. 177:323—340.

    Dyck, J. 1985. The Evolution of Feathers. Zool. Script. Vol. 14(2):137—154.

    Zelenitsky, D.K., Therrien, F., Erickson, G.M., DeBuhr, C.L., Kobayashi, Y., Eberth, D.A., Hadfield, F. 2012. Feathered Non-Avian Dinosaurs from North America Provide Insight into Wing Origins. Science. Vol. 338:510—514.

  6. Great article.

  7. As you said:
    “Only once inside Maniraptora do these filaments start to show a high degree of similarity to bird feathers. Hence why I have no qualms with calling Maniraptoran filaments protofeathers”

    And in fact, the maniraptors (that are categorized as non-paraves maniraptors) are secondarily flightless members of Paraves.

  8. Perhaps a clarification is in order. There are actually two different scenarios that make it possible for the filaments in pterosaurs and ornithischians to coexist with scales (of some kind) and still be homologous to feathers. I have discussed both in my comments, but I might have not always been careful enough to distinguish between them:

    Scenario 1: The ancestral ornithodiran was feathered all the way to the pedal claws. Many of its descendants subsequently re-evolved scales in various body regions; these scales were analogous to the scuta and reticula of modern birds in that they were secondarily derived from feathers, and they were therefore not homologous to the scales of any other sauropsids. Birds belong to a lineage that only re-evolved scales on the lower leg.

    Scenario 2: The ancestral ornithodiran was in the middle of a process where the tuberculate scales of early archosaurs (presumably homologous to those of living crocodiles) were gradually being replaced with feathers. Some of its descendants went further with the replacement than others, giving rise to the diversity of integument distributions we observe in the fossil record. In the bird lineage (not later than in basal paravians but probably as early as in basal coelurosaurs), the replacement became total, resulting in animals that were covered in feathers from head to toe. The re-evolution of scales on the tarsometatarsus that occurred at the base of Ornithuromorpha was a unique event.

    As far as I can tell, both hypotheses are consistent with developmental data, including the evidence suggesting that feathers evolved as the result of an alteration of scale development (what you keep describing as “hijacking” the scale developmental pathway). Scenario 2 just assumes that the alteration or “hijacking” did not affect all body regions at the same time. They are also compatible with the wealth of data showing that feathers are the default integument for the whole body surface of birds and no amount of manipulation — not even dermo-epidermal recombinations — can force their epidermis to produce scales instead.

    If we assume that filaments homologous to primordial feathers were ancestral to Ornithodira and that multiple dinosaur species re-evolved scales en masse using a developmental mechanism that appears to be nonexistent in birds, then we wind up making a level 3′ inference for the ancestral character state (a structure present in the extinct taxon that doesn’t leave an osteological correlate and is not found in the extant relatives). It basically just special pleading at that point.

    If the hypothesis really required an unknown ad hoc mechanism that is completely absent in living birds, it would obviously be very problematic, but fortunately, that’s not the case. It’s quite plausible that the same developmental processes that suppress feather development on the feet of birds (e.g., the Engrailed-1 inhibition of Shh expression described in Prin & Dhouailly 2004) simply acted on other body regions as well in some nonavian ornithodirans.

    On the other hand, your hypothesis (multiple origins of filaments) does seem to rely on an unknown mechanism. It assumes that out of the six times that filamentous integumentary appendages evolved in amniotes, five times it happened within Ornithodira. (The spatulate setae on the adhesive lamellae of geckos and the elongated dorsal scales of iguanas probably aren’t “filamentous” in the same sense that mammal hair and ornithodiran feather-like structures are.) Barrett seems to agree this distribution requires an explanation when he speculates about a “genetic trait that made it easy for [dinosaur] skin to sprout filaments, quills and even feathers”.

    If there was only a partial replacement over time””as has been suggested for feather evolution before””then why would birds need to re-evolve scales again? Why not just “turn off” feather development and “turn on” scale development? Developmental data strongly suggest that birds don’t do it this way because they can’t. Avian-scale epidermis also doesn’t understand the scale-forming signal of lizard dermis (Dhouailly 1975). This suggests birds lost the original scale-forming instructions altogether. If those instructions were “hijacked” to form feathers instead, it would make sense why bird skin can’t form reptile scales.

    Is there any evidence suggesting that the “hijacking” of scale-forming instructions by feathers must have affected the ancestral scales all at once? If so, it would obviously deal a blow to Scenario 2, but it’s hard to imagine that evo-devo research on living birds could conclusively prove something like that.

    Well Microraptor is (or was) considered more basal than Sinornithosaurus (though they might be considered sister-taxa now). Anchiornis used to be more basal than that, though I guess it is considered a avialan this week. That Confuciusornis also shows extensive feathering and is definitely more apical than dromaeosaurs suggests that birds were doing the feathers everywhere thing for quite a long time before they started to dial things back.

    I don’t think Microraptor was ever thought to be more “basal” (= distant from the crown, I assume) than Sinornithosaurus. Mayr et al. (2005) recovered the opposite result: in their tree, Microraptor was more closely related to Confuciusornis than to the rest of dromaeosaurids, including Sinornithosaurus. Everywhere else, the two taxa seem to stick together, even in phylogenies where traditional deinonychosaurs are massively paraphyletic (e.g., Agnolín & Novas 2013). Likewise, I’m not aware of any analysis that shows Anchiornis as more basal than Microraptor or Sinornithosaurus. Of course, none of this takes away from the fact that the “feathers everywhere” phase was quite long-lasting — particularly if Zheng et al. (2013) are right and the metatarsal feathers of Yutyrannus and Sinocalliopteryx represent the ancestral condition for all coelurosaurs, with their absence in other taxa being just a preservational artefact. (That said, feathers on the pedal phalanges are still unique to Anchiornis and Xiaotingia, although I suppose that the same argument about preservation might apply to them as well.)

    I would argue that the best test for feather-scale non-antagonistic association would be to see if one could get protofeathers to sprout from the embryonic epidermal layers in developing crocodylians (in this case, a filament with a barb ridge). If they can form without interfering with scale formation then that would be stronger evidence for a scale-feather association.

    I strongly disagree. Feathers are an evolutionary novelty of bird-line archosaurs; why should we expect to be able to make them appear in crocodiles (complete with barb ridges, no less)? Even if crocodile embryos could somehow be manipulated into developing filamentous integumentary structures de novo, the results would not necessarily tell us anything useful about feather evolution.

    Keep in mind that Rawles discussed this as not being rare in the chicken breeds that she cited.

    I’m not sure about that; the following paragraph was explicitly not restricted to chickens.

    The only birds that I can think of that routinely have a feather-scale mix on their feet are owls. Their apparent easy association is extremely interesting and I would love to know if the filaments coming between the scales actually are feathers or if they are unique filaments akin to the turkey beard. That reminds me, I have access to a frozen barn owl. I should see if I can get samples from its foot.

    Wow. I’m still a bit taken aback that rather than to allow for coexistence between feathers and scales, you would propose that owls have a unique type of integument on their feet which, despite closely resembling ordinary feathers, actually evolved separately from the feathers on the rest of their body and those of all other birds. Perhaps it would be more understandable if the evo-devo papers you cite consistently presented an extremely strong case for the impossibility of any kind of feather-scale association, but that’s just not the case. The papers themselves show quite clearly that the antagonism between the two does not translate into an all-or-nothing relationship, as various degrees of association (rows of feathers surrounded by reticula, feathers emerging from the edges and tips of scutate scales, intermediate structures where the feather and scale epidermis blend into each other) occur in a number of chicken breeds and can be obtained using relatively simple in vitro experiments. (One could argue that this is not at all unexpected: feather formation is tied to the level of Shh expression, which is a continuous variable, not an on/off switch.) In order to preserve the impression that feathers and scales cannot coexist, you have to discard this line of evidence as well, which you seem to be willing to do for reasons that are not entirely clear to me. Note that if “artificially bred animals tend to show wonky developmental stuff anyway”, this is bad news for the whole field of avian developmental biology, which uses the chicken as the model species of choice.

    (it at least seems to stay a coelurosaur, Andrea Cau’s analysis not withstanding)

    Why “notwithstanding”? What Andrea’s analysis shows is that a carcharodontosaurian position for Yutyrannus requires 34 additional steps compared to the most parsimonious tree, and thus can be rejected at a significance level of 0.007 (i.e., safely).

    If Yutyrannus does stay as a tyrannosauroid then there still would not need to be an extra step as the filament node would just engulf practically all Coelurosauria. People have suggested that before. It would make the scales of tyrannosaurids, secondarily derived.

    And that secondary derivation of scales would be an extra step.

    For one there is not a high degree of structural similarity between the structures seen in some dinosaurs and bird feathers. Pterosaurs, Tianyulong, Kulindadromeus and that Psittacosaurus specimen all feature structures that are similar to feathers only at the most basic level. This is the same level that turkey beards, mammal hair and insect setae are similar to feathers. It’s an extremely ambiguous stage.

    The similarities are probably a bit more extensive than that. (Also ˜— insect setae? Really? Why not throw in grass leaves as well? These are filamentous as well, and if protostomes are relevant to this discussion, I suppose that monocots are too.) The bristles in Psittacosaurus and Tianyulong and coelurosaur protofeathers are inferred to be hollow (unlike the hair of most mammals; the bristles of the wild turkey beard are hollow only distally) based on the presence of a dark midline stripe. Feather follicles (which are absent in the beard of the turkey) appear to be present in Sciurumimus, and Rauhut et al. (2012: Supporting Information: 5) claim there is indirect support for the presence of follicles in Psittacosaurus, since its bristles extend beneath the skin layer. The filaments of Kulindadromeus are close to indistinguishable from those in Sinosauropteryx, down to their exact dimensions (lengths of 10 to 30 mm, widths of ~0.15 to 3 mm).

    BTW, do you know of any recent papers on turkey beards? While Sawyer et al. (2003b) established that they are distinct from feathers, they did not propose any hypothesis about their evolutionary origin.

    Coelurosauria is where most of the shared similarities between birds and dinosaurs start to show up, including the formation of protofeathers. Since we are interested in when bird feather development first started, using the animals that first show signs of unambiguous feathers would seem like a logical start.

    That’s circular reasoning — the point where feathers start to show up is part of the question, not an independent fact that can be used to bolster your decision to restrict the prediction to coelurosaurs. It’s also strange that just two paragraphs before this one, you made it clear that you don’t regard the filaments in Sinosauropteryx (an early coelurosaur) as unambiguous feathers, but now you seem happy to accept them as such.

    Chiappe and Gohlich argued that they found “extremely thin filaments at the edge of the preserved soft-tissue.” These “filaments” were so small that even under high magnification they could only be partially observed. Their figure 25 is supposed to show them, but all we get is a blurry image. If these are protofeathers, then they are the thinnest protofeathers yet found. I suspect they are most likely artifacts of either preservation or preparation.

    I don’t get the leap between “they are very thin” and “they probably aren’t feathers, just artifacts”. The authors do not mention how wide the filaments are (only that they are 1.5 to 3 mm long — Göhlich et al. 2006: 23). The fact that the photo we get to see is blurry does not mean that the authors themselves had trouble observing the filaments even at high magnification; in fact, Göhlich et al. (ibid) say they are “particularly clearly visible” ([b]esonders deutlich) dorsal to the 18th to 20th caudal vertebrae and that in several places, it is “clearly recognizable” (eindeutig erkennbar) that they project beyond the outline of the tail. In the absence of further information, it is definitely premature to claim that the filaments in Juravenator would be the thinnest protofeathers known so far. Many of the filaments in Sinosauropteryx are reported to be “considerably narrower than 0.1 mm” (Xu & Guo 2009: 314). If we are supposed to see anything at all in Figure 25, it cannot be thinner than that.

    Repeated placement could just as easily mean that something in the data is drawing Tianyulong towards that node (such as vague characters that are plesiomorphic to many members of the group, which is the problem Kulindadromeus suffers from). Indeed, weak bootstrap and Bremer support values both suggest that there are a couple of characters that are overly-influencing the placement.

    Still not sure what the problem is supposed to be. What are the vague characters that draw Tianyulong and Kulindadromeus away from their right position?

    It would like to know more details on Andrea’s methodology for his study.

    I think the methodology is reasonably clear (except perhaps the exact meaning of the “unknown condition” state); it’s just that there is a lot of room for improvement. However, Andrea apparently plans to run a more detailed analysis in the future, so maybe we’ll soon have not one, but two numerical tests of ornithodiran filament homology.

    The descriptive paper on Yutyrannus did not have strong statistical support at all and was criticized for its matrix choice.

    I was too lazy with the citations, but what I wrote was correct. The two matrices that found Yutyrannus as a tyrannosauroid were the Theropod Working Group (TWiG) matrix used by Xu et al. (2012) and Andrea Cau’s Megamatrice, which was used in two different papers — Godefroit et al. (2013) and Lee et al. (2014). The former was a parsimony analysis of a 992-character version of the matrix; the latter was a Bayesian analysis of the full 1549-character version. Godefroit et al. (2013) recovered both Tyrannosauroidea and a YutyrannusTyrannosaurus clade with Bremer support values of >10, which is what I meant by “considerable statistical support”.

    Interestingly, we don’t know whether or not Xu et al. (2012) had strong support for their placement of Yutyrannus, since they didn’t run the necessary analyses in the first place. Their use of the Theropod Working Group (TWiG) matrix was criticized because it didn’t allow them to test the hypothesis that Yutyrannus was a carcharodontosaur, but that’s irrelevant in this context. Such a position would still mean that tyrannosaurids are bracketed by feathered taxa, and so should be expected to have had feathers as well.

    As I mentioned above, that’s the argument that is typically put forth. However it is not what we actually ever see. Either feathers and scales are both preserved, or they both aren’t.

    I still believe my point about tyrannosaurids is valid (see above). Our best phylogenies strongly suggest that they had feathers, yet scale impressions are all we’ve found so far.

    Also the coarse-grained sediment argument seems less likely now that we have found filament preservation in sandstone (Zelenitsky et al. 2012).

    The discovery of feather traces in sandstone is exciting, but it’s just one more data point. There are hundreds of specimens with feather impressions from fine-grained limestones and just two from a coarse-grained sandstone. That suggests that feather preservation is not necessarily limited to, but far more likely in, finer-grained deposits. Of course, you may still be right if — as Zelenitsky et al. (2012) speculate — feather preservation is relatively common even in coarser-grained sediments and the impressions are often overlooked and destroyed during preparation.

    I contacted Dhouailly about this asking if she had a chance to look at the specimen itself. She does seem okay with calling these structures feathers, but stated that this finding does conflict with current interpretations of feather evolution (including her 2009 study). However she felt more comfortable rewriting what we currently know about feather evo-devo rather than suggesting that these were convergent structures. For that I’m not sure why. One thing to keep in mind is that she doesn’t do work on dinosaurs normally, so the large gulf between protofeathered coelurosaurs and filamented ornithischians (with lots of scaly guys in between) was probably not something she was aware of. I never got a response regarding a first-hand look at the specimen. However she did say that she is working on a follow-up to her 2009 article on feather development.

    That’s great, many thanks for sharing this! For what it’s worth, the website of the Palaeobiology and Biodiversity Research Group has a nice article on Kulindadromeus which says that “[t]he feathers were studied by Dr Maria McNamara and Professor Michael Benton of the University of Bristol, who has also worked on the feathers of Chinese dinosaurs, and Professor Danielle Dhouailly of the Université Joseph Fourier in Grenoble, France, who is a specialist on the development of feathers and scales in modern reptiles and birds”.

    Refs:

    Agnolín FL, Novas FE 2013 Avian Ancestors: A Review of the Phylogenetic Relationships of the Theropods Unenlagiidae, Microraptoria, Anchiornis and Scansoriopterygidae. Dordrecht, Heidelberg, New York, London: Springer Verlag

    Godefroit P, Cau A, Hu D-Y, Escuillié F, Wu W-H, Dyke GJ 2013 A Jurassic avialan dinosaur from China resolves the early phylogenetic history of birds. Nature 498(7454): 359—62

    Göhlich UB, Tischlinger H, Chiappe LM 2006 Juravenator starki (Reptilia, Theropoda), ein neuer Raubdinosaurier aus dem Oberjura der Südlichen Frankenalb (Süddeutschland): Skelettanatomie und Weichteilbefunde. Archaeopteryx 24: 1—26

    Mayr G, Pohl B, Peters DS 2005 A well-preserved Archaeopteryx specimen with theropod features. Science 310(5753): 1483—6

    Prin F, Dhouailly D 2004 How and when the regional competence of chick epidermis is established: feathers vs. scutate and reticulate scales, a problem en route to a solution. Int J Dev Biol 48: 137—48

    Rauhut OWM, Foth C, Tischlinger H, Norell MA 2012 Exceptionally preserved juvenile megalosauroid theropod dinosaur with filamentous integument from the Late Jurassic of Germany. Proc Natl Acad Sci USA 109(29): 11746—51

    Zheng X-T, Zhou Z-H, Wang X-L, Zhang F-C, Zhang X-M, Wang Y, Wei G-J, Wang S, Xu X 2013 Hind wings in basal birds and the evolution of leg feathers. Science 339(6125): 1309—12

  9. Do you take into account the fact that avian scales are essentially modified feathers?

    dinogoss.blogspot.com/2013/09/youre-doing-it-wrong-dino-foot-scales.html

    Kulindadromeus’ integrument is notoriously poorly described, but I think it is believable that it represents an intermediate stage between “feathers” and reticulate scales.

    • Yeah, I’m aware of Matt’s post on the subject (frustratingly, he seems to have gotten most of his information on the feather/scale thing from an earlier post that I had written on the subject). I disagree with his interpretation that most dinosaur scales were reticulae. Reticulae are only superficially scale-like. Even then they most closely resemble the keeled scales on the feet of some lizards. That is quite different from the array of scales observed in dinosaur skin impressions.

  10. Avatar Herman Diaz
    Herman Diaz says:

    What’d be cool is if Kulindadromeus & Tianyulong (& thus, heterodontosaurids) turned out to be primitive cerapods/ornithopods/marginocephalians: Correct me if I’m wrong, but I’m thinking that would mean that ornithiscian quills/fuzz are most probably 1) specific to cerapods in particular (as opposed to ornithodirans/dinos in general), & 2) modified scales (as opposed to proto-feather/true feather homologues).

  11. It occurred to me today that even if the filaments on Tianyulong and Kulindadromeus are truly homologous to proto/feathers, that would actually disprove the common “phylogenetic bracketing” assumption that these structures would be something inherited by all dinosaurs in a given clade—as long as you contrast them with the other ornithischian data. If all the other, more derived ornithischian dinosaurs (e.g. hadrosaurs, ceratopsians, ankylosaurs, etc.) have yielded EXCLUSIVELY scaly impressions unlike these two basal taxa, it would demonstrate that these filaments could be completely lost and supplanted with scales despite what the common appeals to phylogenetic bracketing claim. This of course might have implications for the integument of tyrannosaurs and other theropods. What are your thoughts?

    • Yeah, the presence of a filamented animal next to a scaly close relative, would be strong evidence for convergent (or at least parallel) evolution of those structures. This is one of the lines of evidence that has been used to show that viviparity evolved over 100 times within squamates. In that group there are numerous instances of a viviparous (or ovoviviparous) species being the sister to a closely related, oviparous species. This is also the line of evidence that Barrett et al. 2015 used in their “Scales are the likely default for dinos” paper. That’s another paper that I really need to address, as the paleoblogosphere seems to have given it quite the (IMO, unfair) thrashing.

      Incidentally, a new paper came out today (Mayr et al. 2016) that looks closer at that one Psittacosaurus specimen with the alleged bristles. Their conclusions also strongly warn against homologizing all of these filaments with feathers. The tide may just be finally turning.

      Refs

      Barrett, P.M., Evans, D.C., Campione, N.E. 2015. Evolution of Dinosaur Epidermal Structures. Biol Lett. 11:20150229.

      Mayr, G., Pittman, M., Saitta, E., Kaye, T.G., Vinther, J. 2016. Structure and Homology of Psittacosaurus Tail Bristles. Paleontology. doi:10.1111/pala.12257.