Feathers and filaments of non-avian dinosaurs, part I


By now most people know that feathers are no longer unique to birds. Thanks mostly to a series of wonderful fossils from the Lower Cretaceous Yixian Formation of Liaoning Province, China, we now know that feathers first appeared in non-avian theropods, and were - later on - inherited by early birds...

What doesn't seem so widely recognised is that several researchers had been predicting the presence of feathers in non-avian theropods for an awfully long time, it's just that these people had been mostly dismissed as cranks indulging in way too much speculation. As it happens, the logic they used was fairly sound: bird-like maniraptorans (such as dromaeosaurids) are highly similar to basal birds (such as archaeopterygids), ergo the two probably shared the same sort of feathery integument.

Anyway, while the existence of feathers in non-avian theropods is now well known, a few interesting areas are hardly ever discussed: exactly how were the feathers arranged on the bodies of these dinosaurs, and what did these animals look like when they were alive? While, as usual, there is still an awful lot that we would like to know and don't, the fossils are so good that they have provided us with a great deal of information on the life appearance of the respective taxa.

So far as we know at the moment, true feathers were only present in maniraptorans: the group of coelurosaurian theropods that includes the birds and all the particularly bird-like coelurosaurs (the oviraptorosaurs, therizinosauroids*, deinonychosaurs and, probably, the alvarezsaurids). More simple integumentary structures were present elsewhere in coelurosaurian theropods, including in compsognathids and tyrannosauroids. We still don't know as much about these structures as we'd like to. They are not restricted to the dorsal midline as some have argued but are distributed over the belly, the sides of the head, body and limbs, and they extended down the hindlimb as far as the ankle. They appear to have been hollow, round in cross-section, quite coarse for the size of the animal, and preserved in positions which suggest that, in life, they were pliable. They vary in length according to their position on the body, with the filaments of the compsognathid Sinosauropteryx being 13 mm and 21 mm long on the neck and shoulders respectively. In contrast, the structures on the skull are about 5 mm long, those on the arm about 2 mm long, and those on the distal part of the tail about 4 mm long (Chen et al. 1998, Currie & Chen 2001).

* For a previous article on therizinosauroids see Therizinosauroids and Altangerel Perle.


Because the structures tend to be clumped together it is difficult to be sure of an individual filament's morphology. They might have been simple and unbranched, but Currie & Chen (2001) thought that the structures on Sinosauropteryx might be branched and rather like the feathers of birds that have short quills but long barbs. The similar structures of Dilong - shown in close-up in the image here [borrowed from sinofossa] - also appear to exhibit a simple branching structure (Xu et al. 2004).

What were alleged to be similar filament-like integumentary structures are also present on the Spanish ornithomimosaur Pelecanimimus polyodon, and indeed at least one published life restoration depicts Pelecanimimus as having been covered in the same sort of quill-like structures as are present on Sinosauropteryx and Dilong (Buchholz 1997). However, a brief 1997 report that described soft-tissue mineralization in the Pelecanimimus holotype (Briggs et al. 1997) has been taken by most workers as the definitive last word 'demonstrating' that integumentary fibres were absent from this taxon.

In fact Briggs et al. did no such thing: they described soft-tissue preservation seen in one small patch of tissue, and the absence of integument here doesn't tell us much about the distribution of integument on the live animal. This might explain why a few theropod workers (notably Paul Sereno and Kevin Padian) have continued to indicate the presence of filamentous integumentary structures in Pelecanimimus. Feduccia et al. (2005) argued that Pelecanimimus possessed scaly arms and figured some unusual rhomboidal structures in an effort to demonstrate this. The objects that they figure do not resemble scales and it remains to be seen whether they are anything to do with the integument of this dinosaur. We are still waiting for the full descriptive monograph on this theropod, and when that appears we will finally have more information on this subject.

The big question concerns the homology of the filamentous structures present in these theropods. There aren't any good reasons to consider them as internal collagen fibres, despite arguments from some corners that this is what they are. One could err on the side of caution and interpret them simply as integumentary structures of uncertain homology, and this is perhaps not unreasonable. However, the robust phylogenetic evidence showing that fully feathered maniraptorans evolved from filament-bearing basal coelurosaurs provides good support for the idea that the filaments really are structural antecedents of true feathers.

i-c2bdedc86dea59b4e6fc0606dce511b8-Shuvuuia amnh.jpg

Indeed models of feather evolution have proposed that the earliest prototype feathers were hair-like filaments (Prum 1999, Griffiths 2000, Prum & Brush 2002) similar to the structures of Sinosauropteryx and Dilong: in Prum's model of feather origins, hollow quill-like integumentary structures of this sort were termed Stage 1 feathers. The idea that feathers started out like this also supports Alan Brush's idea that feathers are evolutionary novelties, and not derived from scales. However, the ultimate test of the homology of Stage 1 feathers would be to determine their proteinaceous content: unlike the epidermal appendages of all other vertebrates, feathers are almost entirely composed of beta-keratins (as opposed to alpha-keratins) and, more specifically, they're formed from a group of beta-keratins called phi-keratins [adjacent image of AMNH model Shuvuuia from here].

As yet no one has analysed the Stage 1 structures of Sinosauropteryx or Dilong in order to test their proteinaceous composition*, but tiny filamentous structures discovered adjacent to the bones of the alvarezsaurid Shuvuuia have been tested in this way and, surprise surprise, the structures were composed of beta-keratin (Schweitzer 2001). Alvarezsaurids have been of controversial phylogenetic position but are generally agreed to be basal members of Maniraptora. We can therefore say that beta-keratin-based protofeathers had evolved at the base of this clade at least.

* Paul Davis has repeatedly told me that he would be able to do this if he had a sample. Paul is nowadays a curator of palaeobotanical specimens and probably doesn't have protofeathers on his mind.

i-2506eebd252fe8bdc34e3c6eaca83a64-NGMC 91-A.jpg

As mentioned, while basal coelurosaurs possessed these apparently hollow quill-like 'Stage 1' filaments, they lacked the more complex structures seen in maniraptorans. As we'll see later, maniraptorans possessed vaned feathers with barbs, barbules and hooklets just like those of modern birds. However, their bodies were not covered in vaned feathers as are those of the majority of living birds: instead, it seems that they were at least partly covered in the more simple structures that they had inherited from basal coelurosaurs like Sinosauropteryx. In fact this condition may have been retained all the way up into basal birds: despite all those life restorations clothing archaeopterygids in vaned breast, belly, throat and neck feathers, it seems that their bodies also were at least partly covered in the more simple filamentous structures. The Berlin Archaeopteryx appears to preserve such structures on the back of the neck though pennaceous* vaned feathers were present on its back, at least (Christiansen & Bonde 2004). Vaned feathers were definitely present on the limbs, but we'll come to this later [adjacent image of the feathered deinonychosaur NGMC 91-A from here].


Among other feathered theropods, several of the feathered deinonychosaurs possess unbranched fibres in additional to more complex branched and tufted structures. This is the case in NGMC 91-A, the Sinornithosaurus-like theropod informally dubbed Dave (Ji et al. 2001). Protarchaeopteryx is well known for its fan-like array of 12 rectricial feathers, but it also seems to have sported simple filament-like structures elsewhere on the tail [adjacent Protarchaeopteryx image from here]. Plumulaceous* feathers are preserved in the chest region and tail base, and are also preserved adjacent to the femora (Ji et al. 1998). The basal therizinosauroid Beipiaosaurus also seems to have possessed simple filamentous structures similar to those of Sinosauropteryx (Xu et al. 1999). It's plausible and likely that plumulaceous* feathers were present in places on this animal, and the presence of a pygostyle-like structure provides strong evidence for the presence of a rectricial fan (more on rectricial fans later).

* 'Plumulaceous' essentially means 'soft and downy': it is the opposite of 'pennaceous', a term meaning 'firm and blade-like'.

I have to stop there; more on this subject later. How were vaned feathers distributed on the bodies of maniraptorans, and what might these feathers mean for their behaviour? How widely distributed were 'Stage 1' structures: were they unique to coelurosaurs, or did they actually evolve much earlier? All that and loads more soon.

Some of you might have noticed that this post disappeared and reappeared on March 1st. That was my fault, and it won't happen again.

Refs - -

Briggs, D. E., Wilby, P. R., Perez-Moreno, B., Sanz, J. L. & Fregenal-Martinez, M. 1997. The mineralization of dinosaur soft tissue in the Lower Cretaceous of Las Hoyas, Spain. Journal of the Geological Society, London 154, 587-588.

Buchholz, P. 1997. Pelecanimimus polyodon. Dinosaur Discoveries 3, 3-4.

Chen, P.-J., Dong, Z.-M. & Zhen, S.-N. 1998. An exceptionally well-preserved theropod dinosaur from the Yixian Formation of China. Nature 391, 147-152.

Christiansen, P. & Bonde, N. 2004. Body plumage in Archaeopteryx: a review, and new evidence from the Berlin specimen. C. R. Palevol 3, 99-118.

Currie, P. J. & Chen, P.-j. 2001. Anatomy of Sinosauropteryx prima from Liaoning, northeastern China. Canadian Journal of Earth Sciences 38, 1705-1727.

Feduccia, A., Lingham-Soliar, T. & Hinchliffe, J. R. 2005. Do feathered dinosaurs exist? Testing the hypothesis on neontological and paleontological evidence. Journal of Morphology 266, 125-166.

Griffiths, P. J. 2000. The evolution of feathers from dinosaur hair. Gaia 15, 399-403.

Ji, Q., Currie, P. J., Norell, M. A. & Ji, S. 1998. Two feathered dinosaurs from northeastern China. Nature 393, 753-761.

- ., Norell, M. A., Gao, K.-Q., Ji, S.-A. & Ren, D. 2001. The distribution of integumentary structures in a feathered dinosaur. Nature 410, 1084-1088.

Prum, R. H. 1999. Development and evolutionary origin of feathers. Journal of Experimental Zoology 285, 291-306.

- . & Brush, A. H. 2002. The evolutionary origin and diversification of feathers. The Quarterly Review of Biology 77, 261-295.

Schweitzer, M. H. 2001. Evolutionary implications of possible protofeather structures associated with a specimen of Shuvuuia deserti. In Gauthier, J. & Gall, L. F. (eds) New prespectives on the origin and early evolution of birds: proceedings of the international symposium in honor of John H. Ostrom. Peabody Museum of Natural History, Yale University (New Haven), pp. 181-192.

Xu, X., Norell, M. A., Kuang, X., Wang, X., Zhao, Q. & Jia, C. 2004. Basal tyrannosauroids from China and evidence for protofeathers in tyrannosauroids. Nature 431, 680-684.

- ., Tang, Z.-l. & Wang, X.-l. 1999. A therizinosauroid dinosaur with integumentary structures from China. Nature 399, 350-354.

More like this

First of all, many many thanks to everyone (both here and at SV-POW!) for the congrats regarding baby Emma. I am, shall we say, a little tired right now, but as you'll know if you've visited SV-POW! this morning, the good news as goes blogging is that I've been able to get lots of stuff prepared…
In the previous post we looked at the feathers and filament-like structures that covered the bodies of coelurosaurian theropods. While basal coelurosaurs - compsognathids and tyrannosauroids - possessed filament-like 'Stage 1' structures alone, members of Maniraptora (the coelurosaur clade that…
I used to receive random unsolicited emails from an individual who strongly promoted the idea that birds could not not not not be dinosaurs, that the entire dinosaur family tree was screwed up beyond belief, that 'dinosaurs' had evolved from random assorted diverse archosaurs, that cladistics was…
In 1997, Polish palaeontologist Gerard GierliÅski worked together with artist Marta Szubert to create the amazing furry green(ish), life-sized Dilophosaurus model you see here (and below), nicknamed Dyzio. It was displayed at the Polish Geological Institute's Geological Museum at Warsaw and, while…

Excellent! Thank you for the post and the pics. It's nice to be able to read some real science, instead of some DI or creo "sciency" press release.

What Juravenator can tell us about the distribution of feathers? If it is a very basal coelurosaur (and not a "compsognathid"), outside the "compsognathid-tyrannosauroid-maniraptoriform" clade, then the presence of 'Stage 1' structures could be a synapomorphy of a sub-clade of Coelurosauria.

Very good article. Thanks.

Informative post, but lacking in originality or personal thoughts. This is all information that's been around since 2003.


He's just jealous because you published first. :)

There's also the chance he just got dumped on by his dad, and he's looking for somebody who can't fight back.

Love the blog. Check it every day. Am i right in thinking then that the filament protein was well enough preserved that it could have its tertiary structure identified (as beta-keratin)? This is news to me! Why has no-one done any protein sequencing on these things for potential phylogenetic work? Not enough sample? Problems with purification?
A keratin protein sequence phylogenetic tree with birds, dinosaurs, crocodilians, chelonians, mammals etc. would be incredible.

By Ross Barnett (not verified) on 02 Mar 2007 #permalink


Has any work be done on the type of keratins present in the scales on bird's legs? Could theropods have changed the type of keratin they used in their scales, and then derived feathers from them?

(Paul Davis has moved on to be the registrar at the NHM, so I think he's even less inclined to think about protofeathers than he was before.)

Oh noes! A post that doesn't have much in the way of personal thoughts, or original research! So its stuff you know about, lots of people don't and its nice to have a review article summarising the state of play in these things.

By Dave Godfrey (not verified) on 02 Mar 2007 #permalink

Hi Darren:

Just for clarity's sake, we can't say definitively that feathers originated in non-avian theropods as the earliest known examples as you note are Early Cretaceous, quite a bit younger than _Archaeopteryx_. The only non-avian theropod from that horizon which has integument preserved, _Juravenator_ has scales.

But it certainly seems likely. There are some rather odd ichnites from the Triassic of Connecticut that appear to have been made by feathered theropods, including at least one of a possibly feathered theropod's pubic print.


By Bruce J. Mohn (not verified) on 02 Mar 2007 #permalink

A great post on the Yixian dinosaurs.
Made me remember Yixianosaurus, which I believe is a feathered dinosaur specimen that is just a pair of fuzzy arms. Have their been any studies on where this critter might fall?

By Anonymous (not verified) on 02 Mar 2007 #permalink

Bruce: the non-avian maniraptorans that preserve feathers (Caudipteryx and Protarchaeopteryx and so on) are phylogenetically more basal than birds. Ergo, the most sensible interpretation is that feathers were present in maniraptorans before the clade that we call Aves evolved.

The ichnites that you refer to are a different matter entirely and I'm going to discuss them in a later post (think Gierlinski's giant furry dilophosaur).

As for Juravenator, having spent a lot of time with the specimen I am pretty confident that it is not a compsognathid as proposed by Gohlich & Chiappe. More later.

Also relevent to the Juravenator/scales issue is the following from Peyer's recent (2006) redescription of the French Compsognathus- "However, the caudal vertebrae distal to vertebra 13 display patches of uniformly seized, bumpy structures on their lateral sides (Fig. 2H, 8B, C), resembling the tubercles forming the integument of Juravenator (G�hlich and Chiappe, 2006) in their shape, arrangement, and relative size. These scattered patches of integument are present up to the last preserved caudal vertebra in the present specimen."

IF the Daohugou beds are Middle Jurassic, then at least Epidendrosaurus/Scansoriopteryx and Pedopenna pre-date Archaeopteryx.

Hey, Darren--

Cool post, as always (I cruise your site several times a day as an antidote to my life).

Brief editorial note, not really suitable for posting: it's "plumulaceous", not "plumalaceous".

By Nick Pharris (not verified) on 03 Mar 2007 #permalink