Wednesday, March 07, 2012

Pikaia is most primitive vertebrate known

Yesterday I was talking to one of my colleagues and she asked me if I'd heard the latest news about the Burgess Shale. I confessed ignorance so she told me that scientists had just discovered a primitive vertebrate fossil in the Burgess Shale.

Hmmm ... I was aware of possible primitive vertebrates ("Craniates" is a better term) in the deposits from China (e.g. Myllokunmingia) but I'd never heard of a vertebrate fossil in the Burgess Shale so I thought I'd check out the press release.

It's from my university!!! [Pikaia is most primitive vertebrate known]
Researchers from the University of Toronto, the Royal Ontario Museum (ROM) and the University of Cambridge have confirmed that a 505 million-year-old creature, found only in the Burgess Shale fossil beds in Canada’s Yoho National Park, is the most primitive known vertebrate and therefore the ancestor of all descendant vertebrates, including humans.

The research team’s analysis proves the extinct Pikaia gracilens is the most primitive member of the chordate family, the group of animals that today includes fish, amphibians, birds, reptiles and mammals. Their study is based on the analysis of 114 specimens and is published in the British scientific journal Biological Reviews.
The headline is wrong. Pikaia is a chordate but not a vertebrate as the quotations from the researchers make clear. The press release from Cambridge is only a bit better [Humans' ancient ancestor revealed - as a 505 million-year-old 'eel']
“The discovery of myomeres is the smoking gun that we have long been seeking,” said the study’s lead author, Professor Simon Conway Morris of the University of Cambridge. “Now with myomeres, a nerve chord, a notochord and a vascular system all identified, this study clearly places Pikaia as the planet’s most primitive chordate. So, next time we put the family photograph on the mantle-piece, there in the background will be Pikaia."
Furthermore, this really isn't news. Pikaia was featured in Stephen Jay Gould's book Wonderful Life published in 1989. Even then, Pikaia gracilens was thought to be a chordate similar in broad features to the cephalochordate (non-vertebrate chordate) Amphioxus. This classification was attributed to Simon Conway Morris in 1979. The Wikipedia article [Pikaia] points out that this classification was not universally accepted.

The important points are: (1) that Pikaia is a primitive chordate but not a primitive vertebrate and the press release is just dead wrong about that and, (2) this is old news.

BTW, is Conway-Morris right about Pikaia being the oldest chordate? I thought the fossils from China were older and some of those might even be vertebrates. If that's true then Pikaia lived after the divergence of cephalochordates and vertebrates and it's not even remotely possible that it's our ancestor.

There ought to be a new rule about press releases. Each one should have a statement at the end saying the the press release has been read by the authors of the study and they approve its content.

Conway Morris, S. and Caron, J-B (2012) Pikaia gracilens Walcott, a stem-group chordate from the Middle Cambrian of British Columbia. Biological Reviews. Article first published online: 4 MAR 2012 [doi: 10.1111/j.1469-185X.2012.00220.x]


  1. I read Wonderful Life too and so when I heard about this press release I assumed that the paleontologists had been working to improve the identification of Pikaia as a chordate; even I could tell from reading Wonderful Life that the idea was, how should I say, not well supported at that time. And it looks like that was exactly what happened. So "Old news" is a bit harsh, I think.

    As for the rest: from my reading of the press release, Conway Morris and Caron et al. have only provided better evidence that Pikaia was a chordate; I don't see either of them claiming it was the "earliest". That looks like another embellishment by the writer of the press release.

  2. Conway Morris doesn't say that Pikaia is the oldest chordate. He says it's the most primitive known chordate, and he is apparently using a branch-based definition of "chordate": any deuterostome more closely related to a vertebrate than to a starfish. Or, to put it another way, the phylogenetic analysis in the new Biological Reviews paper puts Pikaia as the sister group of all other chordates.

    The press release made the mistake of equating "primitive" with "old". I hope you aren't making that same mistake. Later we can discuss the mistake of equating "the less diverse of two sister groups" with "primitive".

    1. "The press release made the mistake of equating "primitive" with "old". I hope you aren't making that same mistake."

      I did make that mistake. Conway Morris is using a strange definition of primitive if that's what he meant. Pikaia is younger than several primitive vertebrate so by saying that it is "the most primitive chordate" is like saying that hammerhead sharks are more primitive than elephants or that lemurs are the most primitive primates.

    2. Strictly speaking, "primitive" applies to character states, not organisms or species. It is however reasonable to say that an organism that displays many primitive character states is itself primitive, especially if it displays very few derived states, as would for example be the situation in an ancestor, if we could identify ancestors, which we can't. There is no particular reason to suppose that the fossil displaying the greatest number of primitive states should be the oldest, unless you think the fossil record is very good or that we are likely to find and identify ancestors.

      But Conway Morris's use is hardly unusual or odd. It would be very odd indeed if he had said "primitive" and meant "old". When you say "Pikaia is younger than several primitive vertebrate", what exactly do *you* mean by "primitive"? I don't believe you're thinking this through.

    3. I use the word "primitive" to describe ancient organisms. In that sense, it's the opposite of "living" or "extant." Thus, Pikaia is a primitive chordate and Myllokunmingia is a primitive vertebrate.

      I have no problem with identifying a fossil as the oldest primitive vertebrate or the oldest primitive chordate but that relies on dating fossils and not on some arbitrary choice of evolved characteristics.

      I have a problem identifying one extant species as being more "primitive" than another. I shudder, for example, when people suggest that chimps are more primitive than humans or salmon are more primitive than leopard frogs. All extant species have evolved for the same length of time and just because you can identify that some external features look more like the ancient common ancestor does not mean you can label the entire organism as more "primitive." It could be the case that the enzymes of the citric acid cycle are much more "advanced" in the so-called "primitive" organism.

      The Pikaia of the Burgess Shale lived millions of years after vertebrates diverged from the other chordates and so there are plenty of fossil species that are older than Pikaia. If we could examine their genomes we would find that those other species are a lot closer to the common ancestor than Pikaia. It does not make sense to label Pikaia as the most primitive chordate because we know for certain that this isn't true when you take into account all the evolutionary changes that must have occurred in the lineage leading from the real oldest chordate to Pikaia.

      It would be acceptable to say that certain features of Pikaia probably resemble the oldest chordate more than the same features in Myllokunmingia but that's not the same thing as what Conway Morris said.

    4. I use the word "primitive" to describe ancient organisms. In that sense, it's the opposite of "living" or "extant." Thus, Pikaia is a primitive chordate and Myllokunmingia is a primitive vertebrate.

      You can do that, but be aware that the biologists who deal with ancient organisms (e.g. systematists and paleontologists) don't. "Primitive" is just the inverse of "derived". That is, it refers to conditions rather than age.

      You have a point that it's probably bad to talk about whole organisms as primitive. Every organism has derived features *unless* we are talking about the actual ancestor of some other organism, which must have the primitive state for every character by which it differs from the descendant. Since we can't detect ancestors, that's moot. However, it's not extremely bad to say that organisms displaying a great many primitive traits are themselves primitive.

      Watch out for that word "advanced". It has connotations of inevitable progress, and so is avoided these days in preference to "derived", which doesn't. "Primitive" isn't pejorative either.

      Pikaia, though younger than any of the Chengjiang chordates, seems (according to Conway Morris's paper) to have the primitive state for a great many observable characters for which the Chengijiang animals have derived states. It makes some sense to label it as primitive, its presumed autapomorphies notwithstanding. As for the genomes, we have no access to them, but would it help to say "morphologically primitive"? And what you say is in fact the same thing that Conway Morris said, if you understand the usage in paleontology.

  3. From what I can get from the abstract (I don't have full-text access at the moment), SC-M is using the word "primitive" to mean plesiomorphic, not chronologically old. He said Pikaia could have the most stem-ward position (closest to the origin of the clade) among chordates. That is before the cephalochordate-urochordate split. It's possible that the resemblance between Pikaia and cephalochordates is mostly symplesiomorphic.

  4. There are some interesting things about this new description of Pikaia.

    Firstly, it is clearly not a cephalochordate. It has features more primitive than Amphioxus, including potentially external gills and a terminal anus (rather than a postanal tail).

    Secondly, the myomeres appear to be sigmoidal (W-shaped) leading to the possible conclusion that the V-shaped myomeres of Amphioxus are a derived, rather than primitive, condition.

    The new discovery that tunicates are closer to vertebrates than is Amphioxus has freed us from the constraint to look for sessile chordate ancestors. The new interpretation of Pikaia gives us some better insights as to the likely primitive chordate condition, and makes us more aware that Amphioxus is not a "transitional form" but an extant animal with potentially many of its own derived features that don't represent the primitive chordate condition.

    Christine Janis (aka paleobarbie) --- apologies about posting as anonymous but my university google system is blocking me from signing in any other way.

  5. Larry, you are correct. The body of the press release is also wrong. Haikouichthys has vertebrate morphology and if not a vertebrate via the cladistic definition, it is at least more closely related to living vertebrates. And it is older than Pikaia. Thus, it is impossible for Pikaia to be our direct ancestor.

    John W.

  6. ...the most primitive known vertebrate and therefore the ancestor of all descendant vertebrates...

    That's a weird conclusion...