A Blog by Ed Yong

Tianyulong – a fuzzy dinosaur that makes the origin of feathers fuzzier

Blogging on Peer-Reviewed ResearchWhat happens when you find a feathered dinosaur that really isn’t meant to have feathers? That’s the question set by a spectacular new fossil that adds a confusing dimension to the origin of feathers.

The concept of dinosaurs with feathers is no longer surprising. Birds certainly have them and they are now considered to be living dinosaurs. The infamous Velociraptor and its relatives were covered in plumes, which ranged from the simple quills of Sinosauropteryx to the flight-capable plumes on Microraptor‘s four wings. We know about these prehistoric feathers through the beautiful fossil impressions they have left behind, but a new set of impressions may be the most impressive yet.

They were discovered by Chinese scientists led by Xiao-Ting Zheng, who named their new discovery Tianyulong confuciusi, after the museum that Zheng works in and the famous Chinese philosopher. Its small, agile body,  about the size of a cat, was covered in long, hollow filaments that closely resemble the primitive “proto-feathers” (or colloquially, “dinofuzz”) of other dinosaurs. What makes Tianyulong unique is that it is a very distant relative of all these other feathered species.


So far, all feathered dinosaurs are theropods, a group of two-legged and (mostly) carnivorous animals that included Tyrannosaurus and Velociraptor, and indeed, modern birds. The theropods belong one of the two major groups of dinosaurs, the Saurischia. Tianyulong, however, is a clear member of the other dinosaur lineage, the Ornithischia, which include the various armoured, horned, spiked and duck-billed species. This is the first time that anyone has discovered an ornithischian with feather-like structures all over its body.

More specifically, Tianyulong is a heterodontosaur, a group of small plant-eaters that are the most primitive of the ornithischians. Its position in the dinosaur family tree raises big questions about the origins of feathers. If its filaments are related to the proto-feathers of the theropods (which is possible but not certain), they either evolved independently or were derived from filaments that covered the very earliest of dinosaurs.

Judging by the state of its bones, Zheng’s specimen was probably a young adult that measured 70cm in length (and most of that in its tail). Even without the potential feathers, it would be an interesting find, for most other heterodontosaurs lived in Africa during the Jurassic period. Tianyulong, on the other hand, hailed from Cretaceous China, making it a “living fossil” that was removed from other members of its family in both time and space.

On its skeleton, Zheng found three patches of filaments on the underside of the neck, the back and the tail. Unlike the more complex structures of birds and some other dinosaurs, Tianyulong’s filaments were single structures that never branched. They were very gently curved but otherwise rigid – no bends or waves were found. Those on the tail were especially prominent, measuring about 7cm in length (about a tenth of the creature’s entire body) and about half a millimetre in width. Some of them also had a dark stripe down their middle, a classic sign that they were hollow tubes.

Some scientists have argued that other dinosaur proto-feathers are actually fibres of collagen that have come loose from the animals’ skins. That would certainly make them less interesting, but collagen fibres are solid structures; based on the long, hollow nature of Tianyulong‘s filaments, Zheng rejects this explanation. To him, they clearly stuck out from the animal’s skin.


The big question is whether Tianyulong‘s filaments were actually related to the proto-feathers of the theropods. Zheng can’t be sure based on a single specimen, but he notes that there are definitely similarities. Among the theropods, the proto-feathers of Sinosauropteryx were most similar to those of Tianyulong – they were shorter and more slender, but they also didn’t branch. They also have similarities to the feathers recently found on Beipaiosaurus, which were hailed as the simplest yet discovered.

It’s possible that Tianyulong‘s filaments evolved independently from those of theropods. Indeed, no one has found evidence of proto-feathers in the earliest species of theropods, which suggests that the last common ancestor of this group didn’t have them.

The more intriguing idea is that Tianyulong‘s filaments were a direct part of the evolutionary lineage that led to true feathers, which would mean that the common ancestor of saurisichians and ornithischians was fuzzy. It could have had simple filaments that were retained by Tianyulong, developed into true feathers by the theropods, and lost in many other lineages. Zheng thinks that the similarities between Tianyulong‘s filaments and those of Beipaiosaurus supports this idea.

Only one other ornithischian, an early horned dinosaur called Psittacosaurus, had similar structures but its filaments were sparser, more rigid and only found on its tail. Perhaps these too were elaborate versions of some ancestral filament, borne by the earliest dinosaurs some 230 million years ago.


In a related editorial, Lawrence Witmer says:

“Perhaps the only clear conclusion that can be drawn… is that little Tianyulong has made an already confusing picture of feather origins even fuzzier. Such an outcome is common in palaeontology. But the prospects of new fossils, new molecular and imaging techniques, and even new ideas, offer the hope of bringing the evolutionary picture into sharper focus — and that picture may well end up being of fuzzy dinosaurs.”

Reference: Zheng, X., You, H., Xu, X., & Dong, Z. (2009). An Early Cretaceous heterodontosaurid dinosaur with filamentous integumentary structures Nature, 458 (7236), 333-336 DOI: 10.1038/nature07856

Image: Reconstruction by Li-Da Xing

More on dinosaurs:

16 thoughts on “Tianyulong – a fuzzy dinosaur that makes the origin of feathers fuzzier

  1. Big news!!
    Since pterosaurs are probably close to dinosaurs, “fuzziness” might have originated even earlier within the avian total group. Fascinating.
    About the chart — are we sure Archaeopteryx didn’t have filamentous “protofeathers”? Preservational conditions vary between Solnhofen and Liaoning.

  2. We know about these prehistoric feathers through the beautiful fossil impressions they have left behind

    No, they are not impressions. They are chemical remains of the actual feathers, and/or minerals that crystallized around the decaying feathers.
    Incidentally, the feathers of Archaeopteryx are not preserved as impressions either! They are natural casts.

    are we sure Archaeopteryx didn’t have filamentous “protofeathers”?

    On the other hand, some of the feathers of Sinosauropteryx did branch (according to the 1998 paper, I think), and Caudipteryx had outright semiplumes…
    And then, what theropods are there that are known to have been covered in scales? Juravenator doesn’t count (it had scales, but most of its skin covering is unknown). We have a few patches of Allosaurus and Tyrannosaurus, AFAIK, but that might represent reduction or loss — for Tyrannosaurus, this is strongly suggested by Dilong. Most of the skin surface of Carnotaurus is known, and consisted of scales, but that, too, might represent reduction or loss…

  3. No, they are not impressions. They are chemical remains of the actual feathers, and/or minerals that crystallized around the decaying feathers.

    Interesting. Thanks for the clarification David.

  4. Considering the filamentary covering of pterosaurs, who aren’t far off the dinosaur line, then the whole group may well have been “hairy” in a protofeather kind of way. Wow! Does make me wonder just how dinos would’ve looked in colder climes. Could they have been covered in a seasonal down? I’ve had a mental vision of an Albertosaurus covered in white down stomping through the snow for years, and it now seems there’s definitely something to the idea after all. Cool.

  5. Hey Mike… Kris here from SVP/Dinosauricon. And Zach… how goes it? AK Museum of Nat History still functioning?
    Anyway… I just want to add something about the speculations concerning where within the grand scheme of things this type of integument first appeared. I seem to remember a certain someone’s book… published in 1980-something… where he speculated that integumentary structures of one form or another for the purpose of insulation probably originated waaaaay back with the first proto/earliest dinosaurs. It was immediately marginalized as flights of fancy (pun intended). Finds like this heterodont, and the “quilled” psittacosaur, sorta makes his speculations more like genuine predictions huh? That seems to happen a lot to him. Too bad he never gets the credit he deserves. I was just reading some DML posts about this… some even mention Lagosuchians and “proto-feathers”… But wouldn’t you know it? No one has yet to mention the person I’m referring too. I know you know who I’m talking about. :-)
    Oh, and furthermore, there was this paper a few years back… “Evidence for endothermic ancestors of crocodiles at the stem of archosaur evolution”, Physiological and Biochemical Zoology, 2004; 77 (6):1051-1067, where evidence is presented for endothermy, not only for the ancestors of crocs, but endothermy in stem archosaurs themselves. Again, the feathered heterodont fits oh so nicely into that type of picture.

  6. The species name confuciusi is ill-formed. While the Latin genitive of masculine names is often formed with the suffix -i, in the case of the name Confucius — a 16th-century Latinization of the original Chinese name — the genitive -i should be substituted for the nominative -us. I.e., the form ought to be confucii. Biological nomenclature is frequently poorly formed, but this example is especially egregious.

  7. Does anyone else think: “Dinosaur porcupine!”?
    As to the feather origin “problem”… multiple origins of feather-like extensions would not surprise me. After all, feathers are just modified scales (so is mammalian hair!) and (I presume) all dinos had scales. They were just an evolutionary resource available for selection.
    If any variant of a scale conferred some advantage on its bearer, the characteristic would be favored by selection. Any further err… ‘improvements’ to the modification could amplify the variant. (But I preach to the choir…)
    There’s no reason to assume that gliding/flight were the only possible ‘advantages’ of these modifications. Advantages could have have differed greatly among different lineages: insulation; display; camouflage; active defense; aerodynamics.
    Back to my initial sentence: …. porcupine! In this speculation, the feather-like things were quills. The structure described is almost identical to that of porcupine quills, and their distribution on the body is consistent with defense (throat; spine) and counterattack (tail). (Possibly tangential: the porcupine life style arose independently twice in rodents [old-world vs. new-world porcupines] an origin in other groups is credible.)

  8. I agree that the reconstruction looks very porcupine-like, which leads me to what may be a very naive question. Is there any chance that these filaments can be considered a form of hair as opposed to feathers? What are the defining characteristics of hair that are different from proto-feathers? Please enlighten me on this.
    Also, I agree that such modifications could have had several functions. The distribution on the body fits a display scenario as well as a defense scenario.

  9. Fussy-like filaments in ornithischian Tianyulong are very important discovery, which supports my long ago promoted hypothesis on parallel evolution of theropods and birds, and accordingly parallel and independent origination of feathers (see paper of Kurochkin, Parallel evolution…..Entomological Review, 2006, 86, Suppl. 1, pp. S45-S58). Filaments in ornithischians do not mean that the common ancestor of saurisichians and ornithischians had fuzz. Most likely the filamentous integumentary structures have been originated in these two principal dinosaurs lineages parallel, independently, and within the various time.

  10. ‘Since pterosaurs are probably close to dinosaurs, “fuzziness” might have originated even earlier within the avian total group. Fascinating.’
    Yup – but the original character that defined archosaurs will have been gliding plumes, which explains why crocs and birds have embryonic tramlines (not to mention saurischians-ornithischians actually having dermal decorative features as adults) where ancient plumes went – in the same places Longisquama (and Megalancosaurus for that matter) had plumes. A complexity is that many early archosaur lines were warm-blooded so would have benefited from insulation, so insulative plumage would have been an evolutionary attractor. (Obligate bipedality – ie. unstable equilibrium – means you musn’t need to move in a state of torpor, so obl. bipeds must be some kind of warm-blooded.) But yet another complication is that not all warm-bloods always needed insulation. What the early warmish-blooded crocs had is anyone’s guess.
    Kris on GSP: He’s still being plagued on DML by e.g. Hecht trying to insist on positive evidence. Journalists like to find positive evidence but science actually works by negative evidence, though most scientists don’t realise it, and most journalists – and palaeontologists – don’t care. Although it makes sense to think of a scientist as anyone seeking knowledge (of a global sort of nature), when the up-to-date best practice of knowledge development isn’t followed, practitioners can’t really be called scientists.
    ‘multiple origins of feather-like extensions would not surprise me. After all, feathers are just modified scales’
    John Maynard Smith said homology is a concept ripe for burning, and although it isn’t always clear why he felt that way, I can kind of see it with feathers, which shared bits of developmental mechanism to varying extents with different types of scale, and had a complex 100 mill yr history prior to Archaeopteryx.
    Multiple origins for feather-like things wouldn’t be impossible, it’s just that we have a theory of single origin that explains it better (or at least I do).

  11. In my opinion, the tree (taken from Lawrence Witmer’s summary) is a bit misleading, because it shows all species as existing to the present. In this tree it looks as if Tianyulong is a lone outlier. When you put it more basally on the tree, as in the original paper, it looks more in place.

Leave a Reply

Your email address will not be published. Required fields are marked *