Sunday, November 20, 2011

An Alternate History for "Archaeoraptor"

For those who didn't figure it out, yesterday's mystery bird teaser was a bit of a trick question. In fact, the bird, which I restored on a lark when working on other species, never existed at all. It was, in fact, a restoration of the famous fossil chimera, "Archaeoraptor liaoningensis". Two commentors did get it half right, guessing Yanornis, though the long tail would make that identification impossible (as euornithines, Yanornis had short tails with retractable fans of feathers). You'll notice I modified the image a bit since yesterday: Mickey Mortimer pointed out that Yanornis were not fully toothed, and while the premaxilla had teeth, there is a small edentulous anterior portion which may have supported small beaks.

Despite being the most famous fossil forgery to come out of China, the "Archaeoraptor" debacle really is a success story for the peer review process. You can read a summary of the history of the specimen here. The upshot is that several researchers were suspicious of the chimeric specimen to begin with, and though they rather dubiously chose to submit it for publication anyway, their paper was subsequently rejected by two major journals, Nature and Science, before unfortunately being reported on without review in National Geographic magazine.

In fact, it's not surprising that so many researchers were suspicious of the specimen's validity to begin with. At it combines the skull, body and wings of a Yanornis martini (a songlingornithid euornithine) and the tail of a Microraptor zhaoianus (a microraptorian dromaeosaurid), the "Archaeoraptor" specimen would have been a very odd anomaly had it been real, given our knowledge of bird evolution (even in 1999, when it was first revealed to the public). Note that many news reports have stated that the hind limbs of the composite specimen also come from a Microraptor or even a third, unidentified type of bird, Zhao et al. (2002) noted that they also come from a Yanornis martini. Anyway, I broke out Professor Farnsworth's What-If Machine to try to suss out how our current understanding of bird evolution would be different if the composite nature of the original "Archaeoraptor" specimen had never been discovered.

Intuitively, it seems like a bird with the long tail of a dromaeosaurid combined with the advanced wing configuration and well-developed breastbone of an euornithine doesn't make much sense and would have eventually caused a rather extraordinary rearrangement of the bird family tree. To test this idea, I employed the services of phylogeneticist extraordinaire Mickey Mortimer, who had coincidentally just finished coding Yanornis martini and entering it into a combined Theropod Working Group matrix. Mickey very helpfully humored my request and combined the relevant anatomy Yanornis and Microraptor into a single taxon, and plugged the result into a theropod phylogeny which included only other taxa known in 1999.

I'll let Mickey explain the analysis in more detail:

"Okay, so what I've done is taken my Theropod Working Group analysis and restricted it to taxa described by 1999 (when "Archaeoraptor" was supposed to be described) and coded by the TWG [Theropod Working group - Ed.] by 2005.  I've also limited it to those characters taken from the first TWG analysis (Norell et al., 2001), since I already posted those results on my blog ( and Scott understandably doesn't want more of the Lori results being distributed before publication [Click here for the scoop on Lori, Scott Hartman's unpublished Morrison troodontid -Ed.].  So data-wise, this is equivalent to results I've already posted and with the exception of new specimens of old taxa (like the new Caudipteryx specimens in Zhou et al., 2000), which would be far too tedious to correct for, is representative of our knowledge in 1999.  "Archaeoraptor" is represented by the Yanornis codings mixed with Microraptor tail codings.  Many of those Yanornis codings are from specimens besides the Archaeovolans holotype, but I'm not going to take the time to go through and see exactly which characters can be coded from only that specimen.  In total, 11 codings were changed between Yanornis and "Archaeoraptor".  The entire matrix is 51 taxa and 210 characters."

The result of Mickey's analysis was that "Archaeoraptor" ended up as the sister taxon to Confuciusornis, but with some caveats. First, no other pygostylians (the group uniting Confuciusornis and modern birds)were included, because Confuciusornis is the only pre-1999 taxon coded by the TWG. A complete test of "Archaeoraptor"'s 1999 relationships would require the addition of taxa like Patagopteryx, Songlingornis (which is a potential close relative of Yanornis), Cathayornis, Iberomesornis, etc. Because the only plesiomorphic character of "Archaeoraptor" relative to Yanornis and other euornithines proper are the "longer tail, elongate distal caudal prezygapophyses and rectangular proximal caudal centra", "Archaeoraptor" looks, for 1999 standards, like what we might have expected the sister taxon of Pygostylia to be like. Nevertheless, because the bulk of the specimen is so derived, Mickey reckons the presence of the long tail might not pull it away from euornithines even if other basal pygostylians were included (the long tail would then have to be considered a bizarre reversal).

Mickey also tested a few alternate hypotheses to see how much support they would get. Perhaps the obvious conclusion for a scientists faced with an oddity like "Archaeoraptor" would be that it is probably a sister group to Pygostylia as mentioned above, or somehow intermediate between dromaeosaurids and pygostylians. Mickey tested this by forcing "Archaeoraptor", Confuciusornis, and dromaeosaurs to clade together, which resulted in only three extra steps: this would have been considered a pretty sound hypothesis given the small pygostylian sample. Interestingly, this arrangement would also have resulted in unenlagiines as basal eumaniraptorans and put Archaeopteryx in Troodontidae. Removing "Archaeoraptor" from the mix decreased the likelihood of a Dromaeosauridae+Pygostylia clade by five steps, which Mickey notes is still "plausible" but certainly less likely, and shows that "Archaeoraptor" would have created the kind of link between dromies and more advanced birds touted by the original NatGeo article.

This leads me to speculate that "Archaeoraptor" may have provided "evidence" for the  modular evolution of pygostylians directly from some traditional dromaeosaurs. Modular evolution refers to cases where major traits of a descendant group appear in a taxon which simultaneously retains major "primitive" or plesiomorphic traits of the ancestors taxa. One famous fossil example of this phenomenon is Darwinopterus modularis, which is a long-tailed pterosaur with a generally primitive, "rhamphorhynchoid" body plan, but which has a characteristically pterodactyloid-type skull. Had "Archaeoraptor" been accepted as real, we may now have believed that characteristic euornithine traits evolved first in the skull, forelimbs and torso of taxa which possessed otherwise dromaeosaurid tails (this is almost the exact opposite of how we view bird evolution today, with tail shortening coming very soon after the split between dromaeosaurs and pygostylians). As Mickey found, without more discoveries of basal pygostylian and ornithothoracine birds, our cladograms may have rendered dromaeosauridae paraphyletic with respect to modern birds.

Ok, but how would all this stand up over the years after 1999, with more and better specimens of basal birds from a variety of lineages? From Mickey:

"Here's where I'd expect the inclusion of more pygostylians to have an effect though, since right now with only Confuciusornis, "Archaeoraptor" is effectively the most basal pygostylian.  But if we had an omnivoropterygid, an enantiornithine, and other ornithuromorphs in there, Yanornis' birdy characters would nest it with those and make its dromaeosaurid-like tail a reversal."

So, the discovery of more and better basal bird specimens may have been enough for "Archaeoraptor" to be regarded more and more as a curious side-branch of the avian family tree, less and less relevent to bird evolution as a whole: a euornithine with some rather inexplicable reversals. This would be similar to the way the relevance of the Piltdown Man to human evolution was exponentially reduced by numerous valid specimens before it was finally found to be a hoax.

It wouldn't be DinoGoss without a discussion of taxonomical minutiae, and this is a question that has popped up on the net several times before: if "Archaeoraptor liaoningensis" included the holotype specimen of Microraptor zhaoianus, why isn't the former an objective senior synonym (that is, a name that is synonymous by virtue of being based on the exact same specimen, not a different specimen later assigned to an already named species) of the later?

The crux of the argument is that the original National Geographic article that released the "Archaeoraptor" name did not satisfy the criterion for publication set forth by the International Code of Zoological Nomenclature (ICZN).The article, "Feathers for T. rex", was written by Christopher Sloan for the November 1999 issue of the magazine. While Sloan did technically coin the name "Archaeoraptor liaoningensis" by publishing it in a widely-distributed magazine (and the ICZN does not mandate that names can only be created in peer-reviewed journals), he did not make it clear that he was intending to formally erect a new taxon (usually done by specifying ("new genus and species", "gen. et sp. nov.", or some variation). Furthermore, he explicitly referred to the fact that a formal description of the species was forthcoming. Most observers have interpreted this as falling short of the ICZN requirements for naming taxa, and I don't know of any good reason to disagree.

A more complicated factor is that, among the subsequent publications on the chimeric fossil, some authors did try to formally name the taxon and designate a lectotype (when a type specimen is found to actually represent two or more individuals, a lectotype must be chosen from among them to officially bear the name that originally applied to the lot). Noted BANDit Storrs Olson published an article in a 2000 issue of the Backbone newsletter of the US National Museum of Natural History in which he attempted to remove the tainted name from his own area of study, fossil birds. This is before many BANDits (the "birds are not dinosaurs" crowd) collectively reversed their positions and decided that dromaeosaurs ARE fossil birds after all, thus becoming MANIACs ("maniraptorans are not in actuality coelurosaurs"). Olson therefore designated the tail specimen as the lectotype. This would seemingly make "Archaeoraptor" the official senior synonym of Microraptor.

Not so fast. Olson did not actually describe the specimen or convey intent to coin the name. Like Sloan, he referred the creation of the name to other authors (in this case, Sloan himself, in the mistaken belief that Sloan's article DID coin the name). So Olson effectively specified that the nomen nudum "Archaoraptor" should refer to the tail, but failed to officially create the name, let alone specify its lectotype.

And that is why "Archaeoraptor liaoningensis" remains a nomen nudum; at least, just until some cheeky bastard decides to formally attach the name to something for reals. Technically, nomina nuda are still up for grabs nomenclature-wise...


  1. I agree, and am a completely unbiased source. ;)

    "Mickey Mortimer pointed out that Yanornis were not fully toothed, and while the premaxilla had teeth, there is a small edentulous anterior portion which may have supported small beaks."

    Not only that, Czerkas and Xu (2002) identify a predentary (also O'Connor and Chiappe, pers. obs. in O'Connor et al., 2010) which seems to be standard for basal ornithuromorphs.

  2. Dang, I got so close! I managed to eliminate things down to "Yanornis with wrong tail" even (with help from others on the Hell Creek chatbox), and when you mentioned "half right" I figured it'd be "Archaeoraptor", but to comment then might have bordered on cheating.

    Fascinating post as usual!

  3. stewmorg,
    hey alb i did say it had a tail, and i was half right as well. but my 1st thought was jeholornis

  4. Man, wonderful post. So close and yet so far, very interesting though