A new feathered dinosaur suggests that most dinosaurs had feathers

July 27, 2014 • 8:05 am

What better evidence that birds arose from dinosaurian reptiles than the discovery of a fossil with both scales and feathers? Further, the fossil comes from the right time period: after reptiles had already evolved but before we see modern flying birds with fully-developed feathers.

Of course, we already knew that birds are the only living descendants of dinosaurs—some biologists classify them as dinosaurs—but as we go earlier and earlier back into the evolution of dinos, we’re beginning to find that many, perhaps most, had feather-like structures. That is what we call a “preadaptation”—a feature that could be co-opted later for a different useful function: in the case of birds, gliding and then flight. (“Preadaptations,” of course, didn’t evolve because they’d be useful in the future, for natural selection doesn’t anticipate future needs; it produces features that enhance reproduction in the here and now. But those features can be hijacked for other things later, like penguins’ vestigial wings that became modified for swimming.)

The earliest feathers, as we’ll see on the specimen I’ll show shortly, are small, filamentous structures that occur along with scales. They were of no possible use for flying or gliding, but they wouldn’t evolve unless they enhanced the animal’s reproduction (or its proxy, survival). What were they for? The authors of the paper we’re discussing today suggest this (my emphasis):

Here we report a new ornithischian dinosaur, Kulindadromeus zabaikalicus, with diverse epidermal appendages, including grouped filaments that we interpret as avianlike feathers. This suggests that all Dinosauria could have had feathers and that feathers arose for purposes of insulation and signaling and were only later co-opted for flight.

So what is the finding? The researchers found a 150 million-year-old fossil of an ornithischian dinosaur in Siberia. The beast was about 1.5 meters long (4.5 feet), so it wasn’t large. The important thing is that it was an ornithischian dinosaur rather than a theropod dinosaur, and (I’m going from memory here) all of the decently-preserved feathered dinosaurs that were found previously were theropods. Although there were earlier suggestions that some ornithischians had feather-like structures, the preservation was not nearly as good as that in the new fossil.

The finding is important because ornithischians are in a completely different evolutionary group from theropods, which were saurischians. The former were herbivores and had hips like birds. “Ornithischian” means “bird-hipped,” even though, confusingly, modern birds descended from the other group, the theropods, a carnivorous subgroup of the saurischians. (“Saurishian” means “lizard-hipped.”)

Here’s the dinosaur family tree showing the two different groups and their difference in skeletal morphology. Note that birds (“Aves”) are descended from theropods:

Screen shot 2014-07-27 at 4.47.03 AM

Now if both sauriscians and ornithschians were feathered, that means either that both groups evolved feathers independently, or that they inherited feathers (probably rudimentary filaments that later evolved into full plumage) from their common ancestor. The latter possibility is more likely given the probable developmental complexity of feathers. But even if feather evolution was independent in the two groups, it still suggests that most dinosaurs, not just the theropods whose descendants became birds, had feathers. T. rex (a theropod) could have been covered in down!

But let’s back up and see the details. The new report in Science by Pascal Godefroit et al. (reference and link below, no free download) reports the finding of feather and scale impressions on an ornithischian dinosaur they named Kulindadromeus zabaikalicus (there were actually six skulls and many bones).  The source of the name and conditions of preservation are given by the lead author in a “news and views”-type piece in National Geographic:

The dinosaur’s name essentially means “Kulinda River running dinosaur.” Zabaikalsky Krai is the region of Siberia where it was discovered (which explains its species name, zabaikalicus).

“There were lakes and there were volcanoes there, lots of volcanoes,” Godefroit says. The plant-eating dinosaurs likely died and fell to the lake bottom, where eruptions soon after covered them with a fine ash. That is what preserved the feather imprints with the fossil bones.

“We don’t know how big this fossil bed is, and it is likely we will find more when we go back,” Godefroit says.

The authors of the paper were working in four countries—Russia, Ireland, Belgium, and the UK—showing once again the international character of science. Since I’d like the readers to be able to read at least parts of scientific papers, I’ll put the paper’s abstract below, which is not too technical:


Middle Jurassic to Early Cretaceous [JAC: the Cretaceous went from 145 million years ago to 66 million years ago] deposits from northeastern China have yielded varied theropod dinosaurs bearing feathers. Filamentous integumentary structures have also been described in ornithischian dinosaurs, but whether these filaments can be regarded as part of the evolutionary lineage toward feathers remains controversial. Here we describe a new basal neornithischian dinosaur from the Jurassic of Siberia [JAC: the Jurassic was earlier, from about 201 to 145 million years ago] with small scales around the distal hindlimb, larger imbricated scales around the tail, monofilaments around the head and the thorax, and more complex featherlike structures around the humerus, the femur, and the tibia.The discovery of these branched integumentary structures outside theropods suggests that featherlike structures coexisted with scales and were potentially widespread among the entire dinosaur clade; feathers may thus have been present in the earliest dinosaurs.

 Here’s a figure from the paper showing the reconstruction of the skeleton; the scale lines, which apply to the bones, are 1 cm. (2.54 cm/inch)

Screen shot 2014-07-26 at 3.46.07 AM

The skull, also with a 1 cm scale:

Screen shot 2014-07-27 at 5.44.28 AM

A reconstruction, showing the downy filamentous “feathers,” by the authors:

Screen shot 2014-07-26 at 4.21.54 AM
Reconstruction of Kulindadromeus zabaikalicus. A basal ornithopod dinosaur, with feathers and scales, from the Middle to Late Jurassic of southeastern Siberia. [Drawing by Pascale Golinvaux (Royal Belgian Institute of Natural Sciences)]
And a more colorful reconstruction from the National Geographic piece. The colors are, of course, imagined:

(From National Geographic): This illustration of Kulindadromeus zabaikalicus, a newfound feathered dinosaur, shows it in its natural environment. ILLUSTRATION BY ANDREY ATUCHIN

Here are impressions of scales on the leg (tibia and tarsus):

Screen shot 2014-07-26 at 4.18.01 AM

Large arched scales on the tail (B and C):

Screen shot 2014-07-26 at 4.18.10 AM

Screen shot 2014-07-26 at 4.18.19 AM

Below are the “feathers” on the arm bones (humerus and part of radius and ulna). B. shows enlargement of the white box in “A”, with the filamentous structures growing out of “compound structures”, and C is an interpretive drawing. The authors note:

These occur as groups of six or seven filaments that converge proximally and arise from the central regions of a basal plate. Individual filaments are 10 to 15 mm long. Those on the humerus are wider (0.2 to 0.4 mm) and straighter than those on the femur (0.1 to 0.2 mm). These groups of filaments. . .  resemble the down feathers of some modern chicken breeds, such as the Silkie, which are devoid of barbules.

Screen shot 2014-07-26 at 4.20.43 AM

The fact that feathers appear to be growing out of scale-like features suggests, as biologists have long assumed, that feathers actually evolved from scales, though the authors suggest that the “scales” on birds’ legs and feet are not persistent scales derived from their reptilian ancestors, but evolved back from feathers! Since scales certainly preceded feathers in the fossil record, this shows that truly new structures, certainly involving new genetic information, can evolve (and then be lost, reverting on birds’ feet to scales). That belies the common creationist criticism that new genetic information can’t evolve (we saw that from one commenter earlier today).

Here are some “monofilaments” around the rib cage. These are distributed widely around the head, neck, and thorax:

Screen shot 2014-07-26 at 4.24.12 AM

Enlargement of above (box), showing filaments:

Screen shot 2014-07-26 at 4.24.21 AM

Here’s the money paragraph from the paper:

. . . 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. Further, the ability to form simple monofilaments and more complex compound structures is potentially nested within the archosauromorph clade. . .

Here’s the final statement in the National Geographic article:

“This does mean that we can now be very confident that feathers weren’t just an invention of birds and their closest relatives, but evolved much deeper in dinosaur history,” [Godefroit] adds. “I think that the common ancestor of dinosaurs probably had feathers, and that all dinosaurs had some type of feather, just like all mammals have some type of hair.”

Even so, Godefroit suggests that the largest dinosaurs likely had the fewest feathers, as they wouldn’t have needed them for insulation. “Just like elephants in Africa don’t need fur,” he says.

That suggests that feathers evolved in smaller dinosaurs as insulation, and the largest ones simply lost them, just as elephants, which evolved from much smaller animals, lost their hair (although their mammoth relatives in colder climes either did not, or re-evolved hair). I like the idea that feathers conferred insulation on these creatures, though a signalling function (which means that the feathers probably were colorful, and may have had different colors and patterns in different species) is not out of the question.


Godefroit, P., S. M. Sinitsa, D. Dhouailly, Y. L. Bolotsky, A. V. Sizov, M. E. McNamara, M. J. Benton, and P. Spagna. 2014. A Jurassic ornithischian dinosaur from Siberia with both feathers and scales. 2014.  Science 345:451-455.

61 thoughts on “A new feathered dinosaur suggests that most dinosaurs had feathers

  1. “and (I’m going from memory here) all of the decently-preserved feathered dinosaurs that were found previously were theropods”

    There have been some quilled ornithischians (see Psittacosaurus) but not feathers. It is too soon to say weather these structures are independently evolved from feathers but either way it’s very exciting. You may be interested in these much better images of the specimen’s integumentary structures: http://archosaurmusings.wordpress.com/2014/07/24/kulindadromeus-images/

  2. I like the thought of a downy t-rex. I often found the bird lineage confusing because of the ornithischian line with ornithischian meaning bird and it being named that from having the bird-like pelvis. I always had to remind myself not to think of the name & to remember that theropods were the bird ancestors!

  3. Many of the colorful bird feathers today get their color from mechanical effects, similar to the way that the groves in a DVD make a rainbow. If we ever find a dinosaur feather of that type sufficiently well preserved, we’ll know what color it was….


    1. Except that, since the colour is based on microstructures in the feather and microstructures can be altered by pressure/twisting from sediment above, we still won’t be 100% sure unfortunately. Then there’s the feathers which are coloured by a combination of dietary and/or melanic pigmentation as well as microstructures. It would be amazing to find a structurally coloured fossil feather though.

        1. My link is bad, but see Mark Sturtevant’s reply below, which came in while I was still composing mine, for the correct link.

      1. Structural colours have been found in insects (well, arthropods) in the not too distant past. So finding structural colours in feathers is certainly not incredible.

    2. There are fossils that have preserved tiny granular structures in the feather impressions that look like the pigment bodies (melanosomes) that give bird feathers their color. Different colored melanosomes have different shapes. So the colors and color patterns of a couple dino species has been reconstructed. This finding has become complicated by an argument that melanosomes could have altered shapes during perservation, but the color pattern has not been disputed as far as I know. Anyway, here is an article about that.

  4. Pterosaurs are the sistergroup of the dinosaurs, and may well have some sort of filamentous covering too. That is, feathers might even predate dinosaurs.

    1. I was just thinking that, too. It’s possible that filaments evolved for endothermy, similar to how hairs evolved for synapsids and later mammals, and perhaps pterosaurs were the first to take off with it in the sauropsid line.

      In fact, I might even be daring and suggest that we might see prehistoric crocodile fossils with feather-like structures that they subsequently lost. Admittedly, now we enter a realm of less evidence and more speculation, but it’d be quite a fossil find.

    2. Pterosaurs are the sistergroup of the dinosaurs

      One of several sister groups. The polyphyletic relations of the “reptiles” are poorly resolved.

      Pterosaurs […] some sort of filamentous covering too

      The name “Sordes pilosus” bubbles up from the back of my mind, see here.

  5. “just as elephants, which evolved from much smaller animals, lost their hair (although their mammoth relatives in colder climes either did not, or re-evolved hair).”

    Elephants still have hair, they just do not have very much of it.

  6. Paul Chambers’s wonderfully written pop science book “Bones Of Contention”, about the evolution of birds (now out of print I think), was published around the turn of the century. I remember that at the end of the book there was still some contention as to whether birds had evolved from dinos, although the author clearly favoured that hypothesis. So it seems now that it is well nigh irrefutable and a great illustration of how science progresses.

    1. There are still some hold-outs. Alan Feduccia is a pretty prominent paleontologist who has never really accepted that birds are descended from therapod dinosaurs. His contention is that they are a separate group of related reptiles that evolved in parallel.
      Here is a recent article describing a fossil of a possible Jurassic proto-bird that he interprets to not be a dinosaur. I do not know of the strength of his evidence, but it better be substantial for there to be a paradigm shift.

      1. Czerkas & Feduccia is pretty weak. It depends on a wholesale rejection of phylogenetic methods in favor of arbitrary adjudication of characters.

        1. OTOH, their repeated returning to gnaw on this particular bone does a useful job of keeping other researchers on their toes.
          They’re out on a limb, and I’m pretty sure that they know it.

        1. Thanks, that is very informative. I am not at all surprised to see that the Feduccia paper is not well regarded, given the general silence surrounding it.

        1. For some years (2, maybe 3?), I’ve used the signature line on another forum that –Birds are not dinosaur descendants; birds are dinosaurs, for all useful meanings of “birds”, “are” and “dinosaurs”
          Subtlety has never been my strong point.

            1. I know XKCD well, though I haven’t systematically worked my way through it from one end to the other. So I knew the “feathers” one, but not the Cladistics 1.0.1 one.

  7. What I really want to know is, when Noah brought a pair of these onto the Ark, did he house them in the aviary or on the reptile deck? Perhaps I should get in touch with Ken Ham? I imagine he would know about that sort of thing.

    1. They probably were left behind. Or the fossil is a hoax, b/c ya know there are people in China churning out fake fossils.

  8. Even if these “dinofuzz” structures were not fully homologous to maniraptoran feathers (the basal plate/scle seems different), this find strengthens the inference that the basal dinosaurs were approaching homeothermy — as the big-chicken restoration suggests!

    Probably a [bipedal] step or two ahead of the contemporaneous proto-mammals.

    1. Ah, thanks! I was just wondering how the consensus come down and on what basis, and I see your idea of “not fully homologous … [but] strengthen [an] inference” mirrored in the article palaeosam links to:

      “It cannot be said right now that any of the various filaments seen in Kulindadromeus are genuinely feathers in the sense that they share a single evolutionary origin back at the very origin of the dinosaurs and before the ornithischians split from the theropods and sauropodomorphs. However, at least to a degree that is not actually such a big deal. While the previous filamented ornithischians could at least potentially be written off as oddities, or as ‘only’ having simple filaments, the fact that multiple filament types are turning up in a single species, and that this includes multiple filaments from a single point of origin really make it clear that more structurally complex coverings close to feathers are not exclusive to theropods.”

      That helps me understand how doubts like this can be folded:

      “But Dr Paul Barrett of the Natural History Museum in London, has doubts.

      “Most feathers have a branching structure,” he told BBC News.

      “Instead these look like little streamers coming from a central plate. No bird has that structure in any part of its plumage and none of the developmental models that biologists use to understand the evolution of feathers includes a stage that has anything like that kind of anatomy.””

      [ http://www.bbc.com/news/science-environment-28407381 ]

      And, to be fair, those developmental models have not yet been introduced to the new type of filaments as I understand it, they may handle it well or become better for it. An interesting question in itself, I presume.

  9. And they say there are no transitional forms. They are wrong of course, and this is a fantastic example.

  10. It is significant that scales and feathers are very closely related. Domestic chickens have scaly legs, but an injection of retinoic acid at the right time into the legs of chick embryos causes the development of feathers.

    1. There is also a feathery legged mutation in chickens, and maybe their mutation results in an increase in RA.
      There are also varieties of pigeons with feathery legs, and the leg feathers look a lot like longish flight feathers. A wildly optimistic view would be that these are atavisms of the olde four-winged dinos.

        1. Thanks. If anyone is interested, here is an example. Fantail pigeon.
          Now I want to make clear that there is no reason believe the idea. There is a good chance that the phenotype is just a coincidence.

  11. Despite no longer being a creationist, I feel like there is this lingering creationist part of my brain that likes to point out alleged problems with evolution. Left-overs of brainwashing, I guess. As I read this post, the annoying creationist in my brain pounced on certain phrases like “it seems”, “probably”, “we think”, etc.. “See! There is no evidence for evolution! They don’t *know* anything! They are just guessing!”

    I bet there are a number of ways to deflect this vacuous criticism. Two that come to mind are:

    First, the authors are trying really hard to be honest, and not overstate what they know and to be clear that they are speculating about some things. This is a good thing, not a bad thing.

    Second, creationists conflate “evolution happened” with “how evolution happened.” They pounce on uncertainty regarding mechanism and claim that it reflects uncertainty that evolution happened at all. There is pretty much no uncertainty that evolution happened.

    I’d be happy to hear better / more explanations to improve my ability to explain this to evolution-deniers (including the residual left-over brainwashed irrational part of my brain).


    1. At the risk of being self-aggrandizing, I’d urge deniers (or doubters) to read my book. The massive evidence for evolution is there, and I challenge creationists to explain away that evidence. Of course most religiously motivated evolution deniers won’t accept evidence, for their stand is based not on reason, but on faith.

      1. I’m pretty sure that your book is what helped me understand the second point I made. I should have given you credit in the previous post =)

        You are absolutely right that *most* won’t accept evidence, but for those out there that think trying to persuade a creationist is a complete waste of time, I am evidence that it is not. Thank you for taking the time to write that book. It must have been painful to have to explain things that are probably blatantly obvious to you, but it was the right medicine for me.

        1. Jerry’s book is of course a must-read on this issue. For a more detailed focus on the fossil evidence for evolution I would also recommend Donald Prothero’s “Evolution: What the fossils say and why it matters”.

    2. Nobody can be 100% certain. Using words like “probably”. “it seems” reflects that. In science you do with confidence levels. This is the same for any rational study – you will find the same language used in writing about history, archaeology, literature. You will even see evidence understood this way in crimes (however you will see it argued in more absolutes because the goal is to win a case, not be completely honest). These become weasel words when they are used to obscure instead of illuminate and you can tell the difference from the context.

    3. You are on the right track, and you should give yourself a lot of credit for coming this far.
      You are correct for example to see that scientists use the language of uncertainty when describing interpretations of direct observations(‘it seems’, ‘it appears that..’). But you will see the same language applied in all branches of science, not just evolution. My main field is developmental genetics, and when someone manipulates the expression of a gene and then observes a new phenotype that is what they had expected to see, they will still use the same uncertain language.

    4. The key point is that not knowing for sure is a strength, not a weakness, because it allows you to accept new knowledge and learn from your mistakes.

      Ironclad certainty in the face of contrary evidence isn’t a virtue; it’s just sheer stubbornness. You can’t get to the right answer about anything unless you’re willing to admit the possibility of being wrong.

      1. My definition of science, which is mine, is the apportioning of belief in proportions indicated by a rational analysis of objective observation. There will *always* be error bars, even if the only reasonable alternative is, “Our tinfoil hats have slipped and the aliens are once again controlling our thoughts with their mind rays.”

        Even in physics, where they have more precision in their measurements than anywhere else, they still use error bars. The Higgs, for example, is “only” at a five sigma confidence level; there’s still a 0.00006% chance that it’s a fluke.



    5. I think that evolutionary biologists sometimes leave themselves open to misinterpretation by creationists, IDers, and the like when they use teleological language, such as: “feathers weren’t just an invention of birds and their closest relatives” [my emphasis]

      Clearly, it’s a useful shorthand amongst scientists. But it paints a target on their backs…


    6. Glad to see another person coming on the incremental, asymptotic approach to a correct description of reality.
      There are plenty of good books out there detailing some of the simplest, clearest evidence for evolution. Prof. Coyne’s book is very well spoken of, as is Prothero’s book. It’s not rocket science, nor is it needlessly complicated. Unlike certain other topics, such as theology.

      1. I second those recommendations. Along with the various books from Sean Carroll (the biologist, not the cosmologist this time), along with Your Inner Fish by Neil Shubin. Lots of other good books out there as well.

        1. Sean Carroll (the biologist, not the cosmologist this time)

          One wonders if they’re related. I have this image of an 8-year old game of “Snap” : “I got biology” vs “I just got the rest of the universe.”
          I think I’ve read Shubin’s Inner Fish, but loaned it to someone. I’ve got a closely related “Gaining Ground” by Jenny Clack, which is a bit more monographic and less friendly. And frankly, dull.
          Shubin has recently had a series on UK TV which was probably in the rest of the world. I saw and enjoyed 2 of 3 episodes, but the 3rd has eluded me. Worth watching if you’ve got washing up or ironing to do.

    7. The way I think of the terms of uncertainty used in science (and this can be applied across all sciences) is comparing the confidence levels expressed with a simple process, such as rolling a balanced pair of dice repeatedly.

      Let’s take the odds of rolling snake eyes (a sum of 2 with two six sided dice.) The odds on a given roll are 1 in 36. We can say with roughly 97% confidence on a given roll, “That probably won’t happen.” But now suppose that someone tells you they will roll the pair of dice five times. The odds that they don’t roll snake eyes on one of the rolls is now about 87% (1 – (35/36)^5). We might now say, “it is highly likely that we roll snake eyes.” At 20 rolls, the odds of not getting snake eyes are now 57%, so we might say “it is fairly certain that we won’t roll snake eyes.” This progression continues as the number of rolls are increased until we can say at 120 rolls, “We will probably roll snake eyes at least once” with roughly the same confidence we declared that we probably won’t roll it with a single roll.

      Now, with something like evolution, we may state it as a fact in the same way we’d state it as a fact that you will roll snake eyes if you roll the pair of dice a billion times. Sure, you could roll them a billion times and never get snake eyes, but it would be extraordinarily foolish to declare with any certainty at all that this will happen.

      I am not a scientist, so this is just by view as a lay person and a useful analogy to think about when I read about scientific findings. This sort of certainty actually applies to everything we do in life such as declaring, “I will go to work tomorrow.” In everyday language, I state this as a fact without the countless qualifications I could add such as me getting sick, the train breaking down, getting hit by a bus while I am walking, terrorists striking, etc. In reality, scientific “probablies,” especially those encompassing a robust theory, are many orders of magnitude more certain than the certainty I have that I actually will go to work tomorrow.

  12. Feathers are not moulted the way scales are. In lizards pieces of skin fall off and of course in snakes the entire skin is shed. I don’t know if this would be a peak load on the metabolism or even a risk due to breach of the integument. Could there be an adaptive advantage to protofeathers simply by virtue of the difference in moulting?

    1. It is a puzzle. But since the current find is from the ornithischian dinosaurs, it could be that their way of growing and replacing feathers is different from saurischian dinosaurs. Maybe the feather/filaments are periodically shed separately from the integument of the basal scale. It would be good to see the scales with a little stub of a feather/filament just starting to grow out.

  13. That individual at the back of the Nat Geo illustration looks like it was inspired by a wallaby.

    Not that there’s anything wrong with that…

Leave a Reply