Earliest bird was not a bird? New fossil muddles the Archaeopteryx story

By Ed Yong | July 27, 2011 1:00 pm

The eleven specimens of Archaeopteryx are some of the most iconic and captivating fossils in existence. The fingers end in claws, the tail is long and bony, and the head – arched back in the throes of death – contains toothed jaws. But the splayed arms are lined with the faint but unmistakeable outlines of feathers. This was an animal halfway between a small flesh-eating dinosaur and a modern bird. In fact, Archaeopteryx is widely heralded as the first bird, occupying a pivotal position in the origins of this group.

But Xing Xu from Linyi University thinks that this first bird was nothing of the sort. The Chinese palaeontologist, who has found one fascinating dinosaur after another, has identified a new species called Xiaotingia that threatens to oust Archaeopteryx from its position.

By comparing Xiaotingia’s features with those of Archaeopteryx and other related birds and dinosaurs, Xu has drawn up a new family tree (see slideshow below). In it, Archaeopteryx sits with Xiaotingia among the deinonychosaurs, a celebrity-filled group of small, predatory dinosaurs that includes Deinonychus and Velociraptor. The lineage that led to modern birds perches on a different branch of the tree.

This doesn’t change the fact that birds evolved from dinosaurs – it merely relegates Archaeopteryx to the sidelines of that process. In its place, species like Epidexipteryx and Epidendrosaurus take up the mantle of earliest birds. It is a tentative revision but a bold one (Xu himself admits that the new family tree is statistically weak). “It’s been a good run for Archaeopteryx,” writes Larry Witmer in a related editorial. “This finding is likely to be met with con­siderable controversy (if not outright horror).”


Xiaotingia lived in China during the late Jurassic period, and was about the size of a pigeon. Until recently, its beautifully preserved fossil sat innocuously in the Shandong Tianyu Musuem of Nature, among a thousand-strong collection of feathered dinosaur skeletons. Xu noticed it when he visited the museum. “I immediately recognized that it was something new, but honestly did not expect that the discovery would change the dinosaur-bird family tree.”

Xiaotingia’s presence is crucial. When Xu re-ran his analysis with exactly the same species except for his new discovery, Archaeopteryx was restored to its original position as a proto-bird. This new specimen has made a big difference.

However, it’s not the only fossil to have called Archaeopteryx’s status into question. In recent years, fossil hunters have unearthed several species that are either primitive birds (or close relatives of them) that look very different to Archaeopteryx, including Epidexipteryx, Jeholornis, Sapenornis.

Meanwhile, others have noted that Archaeopteryx has many features that are only found among the deinonychosaurs. For example, they share a distinctive hip bone, and they both have a large hole above their noses (the “premaxillary fenestra”) that other birds and dinosaurs lack. Any many of the features that supposedly characterise Archaeopteryx and other birds, such as feathers, a wishbone and long powerful forearms, are also found in deinonychosaurs.

This is more than just a matter of shuffling cards. Archaeopteryx’s position has been so sacrosanct that its body had guided many of our ideas about the origins of birds. It grounds our understanding of this group. For example, it was previously thought that most primitive birds and their closest dinosaur relatives had lightly built skulls, which might have been useful for flight. “Instead, our study suggests that primitive birds had robust and rigid skulls,” says Xu.

“Our study also suggests that most primitive birds were herbivorous animals,” he says. “Previous studies which suggested that flight evolved in the [meat-eating] context (such as wings evolving for catching prey) need reconsideration.” Witmer writes, “Clearly, without the safety net of good old Archaeopteryx at the base of the birds, we’ve got some fresh work to do.”

But Gerald Mayr, who studies fossil birds at Germany’s Sneckenburg Museum, is unimpressed with the new discovery. “I fear that it is a bit hyped and that the conclusions are not as novel as the authors claim,” he says. Mayr is one of several palaeontologists who think that the deinonychosaurs are actually birds themselves. According to him, they’re flightless members of a group that includes Archaeopteryx and modern birds, like smaller extinct versions of today’s ostriches and emus.

The problem is that all of these reconstructions are weak. Mayr admitted as much about his own model back in 2006, and Xu says that his new family tree only has “tentative statistical support”.

Creationists will doubtlessly pounce upon this story and quote-mine articles for supposedly damning phrases. But revisions and uncertainities like this are to be expected. As with all big evolutionary shifts, there wasn’t a simple linear route from dinosaur to bird. Instead, animals at the time developed a whole range of different body shapes that were eventually tested and winnowed by natural selection. “There were lots of experimental trials”, says Xu.

The fossil record is full of these failed experiments and evolutionary dead-ends, all preserved alongside the success stories that eventually gave rise to modern species. This makes it very difficult to construct a robust family tree. Witmer sums it up brilliantly in the conclusion of his editorial:

“Just as Xiaotingia moved Archaeopteryx out of the birds, the next find could move it back in — or to somewhere else within this fuzzy tangled knot that makes up the origins of birds and bird-like dinosaurs. That said, during this sesquicentennial anni­versary of Archaeopteryx, which is being honoured with exhibits and commemorative coins, the bitter irony may be that it may not have been the bird we’ve always thought it was. But Archaeopteryx will remain an icon of evo­lution, perhaps even more so now, providing compelling evidence that, as we should expect, evolutionary origins are rather messy affairs.”

Reference: Xu, You, Du & Han. 2011. An Archaeopteryx-like theropod from China and the origin of Avialae. Nature http://dx.doi.org/10.1038/nature10288

More on early birds and feathered dinosaurs:


Comments (12)

  1. Eric Murphy

    The chances that Archaeopteryx was actually directly ancestral to birds have always been pretty slim. But even if it were actually true that Archaeopteryx where the direct ancestor of all modern birds—how would we know that? The same is true of any other example in the fossil record.

    No matter how close to the actual node connecting all modern birds a given fossil lies, there is always the chance a later fossil will be closer to that node. And regardless of how close to a node any given fossil is, there is no real way to tell if it actually is AT the node.

    This is why strict cladists have long since giving up trying to identify actual direct ancestors in the fossil record.

  2. So the new guy comes out as a sister taxon to Anchiornis, but in this tree, Anchiornis is not a troodontid. I thought there was strong support for Anchiornis as a troodontid? It’s also strange that the proposed phylogeny posits scansoriopterygians as the new “first birds.” We certainly need more, better-preserved, more-completed members of that group before we can make any solid determination as to their closest kin.

    Still, good to see another Archaeopteryx-like theropod come out of the ground. If nothing else, it shows that Archaeopteryx may constitute a “grade” of bird-like theropods that were more widespread than we usually think.

  3. chris y

    I wonder if there’s a failure of language here. Linnaean classification works, up to a point, because in most cases it’s easy to distinguish taxa at any given level, either by morphology or genetics. But it runs into serious problems in the historical dimension.

    Suppose, for the sake of argument, you define aves as the LCA of Passer domesticus and Archaeopteryx lithographica and its descendents. Then you get into your time machine and go and look at this LCA. You notice that in the next valley lives a species which is almost indistiguishable, being a sibling of aforesaid LCA. In any sane world you would name it as another species within the same genus, but you’ve defined aves in such a way that it can’t be a member. In fact, in traditional taxonomy you’ve put it in a different class.

    This way lies madness, it seems to me. As a total amateur in these matters, I react by thinking that “What was the first bird?” is a bad question, possibly meaningless, certainly unanswerable, even if we had all the data. But the popular press, at least, continues to ask it. What’s a simple soul to do?

  4. Bob LaVesh

    The same is probably true for many of the “known” species “ancestors.”

    It will always be very difficult to determine with a high certainty that one species begot another- we will never know with many if they are true ancestors or just a branch from a common ancestor.

    This in no way diminishes the important part Archaeopteryx has played historically in promoting evolutionary sciences (even if our understanding then was incorrect) – and even as an important clue between the link between dinos and birds.

  5. Karl Zimmerman


    People have suggested this day would come sooner or later – it’s interesting to see it may finally be here. I can’t wait to read Andrea Cau and Mickey Mortimer’s thoughts (once they plug the data into their own Theropod matrices)

    Whether Archie remains a bird, however, depends upon your definition of Aves. It’s only not a bird if your definition is stem based and something like “modern birds and all fossil taxa closer related to birds than Dromeosaurus” or some such. On the other hand, if it’s a apomorphy-based definition, say due to flight, than all Deinonychosauria may come along for the ride.

    One thing he seems to have gotten wrong though is on the origins of flight. Although none of the basal “birds” seem to have been great fliers, there are now taxa in bot major branches of Paraves which seem to be flighted. Thus either flight (or at least proto-flight) evolved twice, or was basal to the entire group, meaning it might have an insectivorous origin regardless.

  6. Hang on a minute. No one said “ancestor”. The first time that word comes up is in comment #1. When I talk about “earliest bird”, that means “earliest member of the bird lineage known from the fossil record”, not “direct ancestor of all birds”. Regardless, the point is that the new family tree says that Archaeopteryx isn’t a bird at all.

    Also, fully agree with Bob LaVesh’s third paragraph: “This in no way diminishes the important part Archaeopteryx has played historically in promoting evolutionary sciences (even if our understanding then was incorrect) – and even as an important clue between the link between dinos and birds.”

  7. amphiox

    Mayr is one of several palaeontologists who think that the deinonychosaurs are actually birds themselves. According to him, they’re flightless members of a group that includes Archaeopteryx and modern birds, like smaller extinct versions of today’s ostriches and emus.

    Looking at the cladograms shown, this doesn’t actually seem that much different a hypothesis. You’re basically just taking Paraves and calling them all birds, rather than calling the Avialae birds and the Deinonychosauria not-birds, aren’t you?

  8. Lars

    I’d like to propose the colloquial name “trouserbirds” for this group. All of the reconstructions I’ve seen make them look as though they are wearing old-fashioned uniform trousers.

  9. “Regardless, the point is that the new family tree says that Archaeopteryx isn’t a bird at all”

    This is only valid with certain definitions of Aves: the ones that Xu et al. prefer.
    If you use Sereno 2005, or Chiappe, 1997, or Padian, 1997, then Archaeopteryx keeps being a member of Aves.

  10. Daniel J. Andrews

    Back in 2000 or earlier, there were discoveries of fossils that had more ‘advanced’ bird features than Archaeopteryx yet were older than Archaeopteryx so my thought was that already people were thinking it was not the earliest known bird (hazy memory alert though).

    I know Written in Stone has a full chapter on this so I’ll have to go back and reread that chapter keeping in mind this new finding.

  11. Eric Murphy

    @Ed Yong: but doesn’t the statement “The lineage that led to modern birds perches on a different branch of the tree” imply that someone, at some point, was claiming Archaeopteryx was directly ancestral to modern birds? If the lineage that led to modern birds did, in fact, branch from Archaeopteryx, that would necessarily mean Archaeopteryx is directly ancestral to Neoaves.

    And as others have pointed out, whether or not Archaeopteryx is a “bird” is a matter of definition: how we define birds.

  12. Great discussion. I have little to contribute to the science except to say that the Nova special I saw on this topic clearly demonstrated what a p***ing contest this kind of research can become when 2 scientists have clearly opposing views which then get into the media. You’d think they’d put their heads together for the good of discovery; although one scientist in particular seemed as though he was all about disproving the other.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Not Exactly Rocket Science

Dive into the awe-inspiring, beautiful and quirky world of science news with award-winning writer Ed Yong. No previous experience required.

See More

Collapse bottom bar