Saturday, August 11, 2012

Did Sinosauropteryx Have "Protofeathers"?

Life restoration of the S. prima type specimen by M. Martyniuk
Over at Jaime Headden's blog The Bite Stuff, Jaime recently wrote a great article about the newly discovered basal coelurosaurian or orionidian Sciurumimus. The article touched on the fact that its feathers were reported as stage 1 (simple, unbranched filaments, often referred to as "protofeathers") in the parlance of feather development researcher Richard Prum. In the comments, Heinrich Mallison pointed out that they *look* like stage 1, but as Foth showed, crushed feathers (even those of modern birds) often look much more primitive than they are due to taphonomic effects. (I've mentioned before in a few places that it's really unfortunate Foth's important paper seems to have been mostly ignored by others writing about fossil feather types). Jaime defended this by saying that Sinosauropteryx, which is more derived, also is 'generally agreed' to have had stage 1 feathers. It's true that this is the general agreement in the literature, and it's also true that the general agreement has been challenged (effectively, in my opinion). I posted a reply to the article, but the fact that this keeps coming up lately has made me think that maybe I should be trying to publicize this more widely, so my comment is reproduced below along with some more commentary on the issue.

Jaime wrote:
"I will generalize at this point that they are modified stage I, based on the likely LESS basal Sinosauropteryx prima and general agreement on which stage it has."

I replied:
Another case of general agreement about fossil feathers that (as with Foth's paper) was demonstrated to be wrong long ago in important research which was subsequently ignored by all researchers who simply accept the meme that Sinosauropteryx has stage 1 filaments without thinking about it too much.

Currie & Chen 2001 (which as far as I know is the only non-BAND paper to examine the S. prima integument in detail and actually publish their findings) concluded that "The mixture of thick and thin strands close to the body, the increased presence of thinner strands distally, the fact that the thicker strands are positioned close to the body and are normally oriented at higher angles from the body than more distal strands (Fig. 12g), the presence of areas where many of the finer strands lie adjacent and parallel to each other (even kinking together in a few places), and the tendency of finer filaments to angle away on both sides from thicker structures (Fig. 13a) all suggest a feather-like structure with central shafts and plumulaceous barbs."

So, the feathers of S. prima are at least Prum's stage 3! So why does everyone continue to refer to them as "protofeathers" or stage 1 feathers? It's not like this is an obscure paper. Sorry if I sound exasperated [but] the the state of the literature on fossil feathers is absolutely atrocious and seems to be plagued by lack of research and plain laziness. Not only do important specimens get barely a sentence mentioning the feathers after pages and pages of detailed osteology, but the papers that do bother to describe feather structure go almost completely ignored by subsequent authors.

Currie, P.J.; and Chen, P.-j. (2001). "Anatomy of Sinosauropteryx prima from Liaoning, northeastern China". Canadian Journal of Earth Sciences 38 (1): 705–727.

You can read the comment thread at Jaime's article for the full discussion. To summarize, Jaime replied that while he's aware of Foth and Currie/Chen's work, he still prefers to default to the stage 1 identification until completely unambiguous evidence is found. Jaime notes that Currie/Chen only demonstrate barbs on a single side of the rachis in the S. prima holotype (but unambiguously on both sides of the rachis in one of the adult referred specimens), which would be bizarre and out of line with Prum's developmental model, unless they were secondarily derived from true stage 3 feathers (IMO this apparent one-sided vane is probably best explained by taphonomy due to the severe level of overlap and impossibility of finding a completely isolated feather). Jaime also notes the possibility that there are in act two distinct types of stage 1 feather present, a short, thick, rachis-like filament and longer, thinner, barb-like filaments. But given the fact that, as Currie/Chen reported, the thinner filaments are almost always at an angle to the thicker ones, the simplest explanation is that these are standard, barbed, plumulaceous feathers similar in structure to modern down.

As I pointed out in a rather convoluted way in my initial comment, there are two main problems with the state of research into fossil feather types. One, feather anatomy often gets the short shrift in descriptions of new fossil specimens, even those that spend a lot of time detailing the intricacies of the bone anatomy. Second, there seems to be a systematic problem in the scientific literature, where researchers (for whatever reason) ignore research that has implications for the conclusions they are drawing about new feathered specimens. For example, the Sciurumimus description states that based on their examination, the feathers in the specimen appear to be monofilaments and therefore stage 1. but it would be important to note that this could possibly be an illusion created by taphonomy, and that in principle taphonomy can makes fossil feathers appear more primitive than they are. This is not to say that the feathers are more advanced than they seem, but that other research suggests they very well could be. This is a pretty major omission. Even more advanced feathers suffer from this problem.  In a 2010 Bite Stuff post on the feathers of the microraptorian specimen "Dave" claimed that the body including arms were covered in branched filamentous feathers rather than vaned feathers, when the paper explicitly states the wing feathers were vaned and pennaceous, and infers the presence of barbules based on the arrangement of barbs.

I don't mean to pick on Jaime here (though I continue to see people citing his article to defend reconstructing Dave with fluffy arm feathers). His reasoning in the Sciurumimus article is fair, if a bit over-conservative in my opinion given the available data. My major complaint is the fact that these issues are never even raised in the literature, even just to shoot them down as not providing strong enough evidence. Instead, these potentially important papers continue to be ignored by relevant research. Whether or not Currie and Chen convincingly demonstrated that the feathers of S. prima are branched, or Foth convincingly showed that taphonomy tends to make feathers look more primitive... Well, to not even acknowledge this stuff in a paper about primitive feather fossils is just poor science.


  1. In the comments Jaime does also link to this close-up of the feathers on a specimen referred to Sinosauropteryx, noting that these do appear to be convincing stage 3 feathers.

    Oddly, the Sciurumimus paper cites Foth (2012) for their comment that the feathers of Sciurumimus may be simpler than they appear (apparent branching being the result of overlapping feathers)... and Foth is also one of the authors on the description of Sciurumimus!

  2. Matt, great summary! I wish to stress that my "Dave" article was written while Foth was working on his research, before it was published. I did not, at the time, consider differential preservation as a strong factor influencing the prservation. If I had another go at it, I would try to distribute the stage III feathers more evenly on the arms, legs, and so forth. My primary reasoning for how they are NOW being anything like how they MIGHT have been is that extremities USUALLY preserve pennaceous vanes more clearly, yet the arms seem to lack them; Foth even backs this up. Body feathers are obscured (as in "Dave") while peripheral feathers appear as filaments (as in "Dave") ... but pennaceous vanes are only on the leg. I think this IS peculiar, and I would adjust this now. In fact, this inspires a new post.

    You are correct: New work demonstrating integument does not preserve as in life implies we should be more clearly speculative of preservation of "stage I" in feathered dinosaurs. But how you distinguish correctly and incorrectly preserved stages I, II and III? That's the kicker.

  3. I completely agree and are guilty of improperly assuming reported stage I feathers are correctly interpreted myself.

    Perhaps in an odd turn of events, BAD supporters are being BAND-like and not thinking about the issue because preserved stage 1 feathers on basal coelurosaurs fits so nicely with Prum and Brush's model of feather development.

    One interesting question is then- are there any definite non-branched feathers preserved? Psittacosaurus, if its tail quills are homologous. What about pycnofibers? Are any pterosaur specimens well enough preserved to rule out taphonomic issues?

    1. The EBFFs of Beipiaosaurus, perhaps?

    2. I would suggest that the EBFFs (elongated, broad filamentous feathers) are also taphonomic (using Foth's PRPF, proximally ribbon-like pennaceous feather, argument for Similicaudipteryx yixianensis. In this model, the "flat, broad" structure is merely and likely always will be the calamus sheath enclosing the filamentous structure inside (rachis and barbs). Preservation issues MAY account for their not being a distally pennaceous structure as in Similicaudipteryx yixianensis for "EBFF" taxa like Epidexipteryx hui or Beipiaosaurus inexpectatus.

      In other words, "EBFFs" do not exist ... they are developmental and preservation artefacts.

  4. Has the term "protofeather" ever really been defined as only including unbranched filaments?

  5. @Brad
    As an informal term usually presented in quotes, I don't think it has any formal definition, but in practice I usually see it referring to stage 1 or 2 feathers. It would be incredibly silly to apply "protofeather" to say, Stage 3. Nobody says baby chickens are covered in "protofeathers."

  6. One thing many people have problems with is that "lack of evidence" is not "evidence of lack".

  7. @Heinrich Mallison
    I'd agree with that in principle, but not in some more complex situations. For example we have enough specimens of basal coelurosaurs (tyrannosauroids and compsognathids) to be able to pretty confidently say, for example, that they didn't have stage 4 remiges on their arms. I think it's not so much lack of evidence but assessing the quality and reliability of the evidence we do have that's the issue here.

  8. Makes me think of when those rumours of "branched feathers" preserved with a _Tyrannosaurus_ were dismissed as "maybe a bird was preserved alongside it" just because branched feathers didn't fit the then current model of feather evolution.
    I smelt bullshit at the time, and I smell it more strongly now.

  9. I didn't remember that Foth's paper is out; all I've seen is the poster he presented at this year's meeting of the European Association of Vertebrate Palaeontologists. Could you provide the reference?

    (And could you shrink your profile pic?)

    rumours of "branched feathers" preserved with a _Tyrannosaurus_

    Never heard of that. Details, please!!!

    1. @David Marjanivoc:
      Foth's paper is

      Foth, C. 2012. On the identification of feather structures in stem-line representatives of birds: Evidence from fossils and actuopalaeontology. Paläontologische Zeitschrift 86:91-102.

      Not sure what you mean about the profile pic...?

      "Never heard of that. Details, please!!!"
      I wish! Have to dig up the original rumor from some message board or other and see who the ultimate source was. Only thing I remember offhand is that there was speculation the specimen represented a post-mortem association (from the Lance I believe) and not a branched feather on T. rex, though only because the original source(s) didn't think this would be phylogenetically plausible. If compsognathids do have Stage III feathers, there's no reason to think it's not real, if and when it's ever described.

  10. Thanks for the info. Never mind the pic, it's fine now.