A New Twist in the Tetrapod Tale

The Devonian fish Eusthenopteron crawls along a mudbank in a diorama at the National Museum of Natural History. Photo by the author.

I love outdated museum displays. They marvelously represent the “history” part of natural history exhibits – dusty dioramas of old ideas that are a baseline for how much our understanding has changed.

One of my favorite displays is tucked away in a dim corner of the Smithsonian National Museum of Natural History. If you just amble through the main gallery – beneath the osteological frames of the ever-popular dinosaurs – you’re bound to miss it. The humble little scene, displayed in a glass case along a wall running behind the Mesozoic celebrities, envisions one of the most important moments in evolutionary history. On a dried, cracked mudflat, an intrepid Devonian fish makes a foray out onto land. The scene represents the first steps – so far as a fish can be said to step – any vertebrate took on into the terrestrial realm, and an artistic rendition of the salamander-like Ichthyostega in the background signifies where that uncomfortable piscine shuffling would eventually lead.

Starring the fleshy-finned fish Eusthenopteron, the Smithsonian scene is a distillation of the romantic idea that the first tetrapods – the four-legged vertebrates that colonized the land, our forebears among them – were forced to adapt to a terrestrial world as arid climates dried up the ponds archaic fish had relied upon. This idea, proposed at the beginning of the 20th century, became the dramatic canon for the origin of tetrapods. There were only two options for any creature trapped in the shrinking mudholes – evolve or die. Dire necessity was the mother of anatomical invention.

But the heroic story of brave fish biting their fins at fate and adapting to a harsh, changing world has faded into a footnote in the history of science. Even though paleontologists were correct that the stacks of bones inside Eusthenopteron fins represented the archaic beginnings of what would become our own arms and five-fingered hands, the ecological backdrop was all wrong. The earliest tetrapods lived in lush, swampy environments that were not in risk of evaporating.

The anatomical evidence played a part in scuttling the traditional story, too. Eusthenopteron would have been useless on land, and recently-discovered “fishapods” such as the roughly 380 million year old fish Panderichthys and the celebrated, 375 million year old Tiktaalik used their modified, limb-like fins to get around in the water. The origin of arms, legs, fingers, and toes wasn’t a transformation that happened on land. Limbs were an aquatic innovation that just so happened to be advantageous when tetrapods began to venture out of the water.

A restoration of Ichthyostega – with the skeleton visible inside – by Julia Molnar.

Even the abilities of Ichthyostega and kin have come into question. With four limbs and differentiated digits, Ichthyostega would seem to be a creature that lived right at the margin between water and land. But the creature’s anatomy doesn’t match our expectations of terrestriality.

When paleontologists Per Ahlberg, Jennifer Clack, and Henning Blom reexamined and reconstructed the skeleton of Ichthyostega in 2005, they found that the overlapping ribs of the tetrapod would have significantly restrained the kind of side-to-side walk expected for such a salamander-like animal. Instead, Ichthyostega might have had a weird shuffle – in which the legs moved but the creature’s torso was kept rigid – or used an “inchworm” kind of crunch. Simply put, the tetrapod was rubbish on land, and the authors concluded that “Ichthyostega appears to be an early and ultimately unsuccessful attempt at adapting the tetrapod body plan for terrestrial locomotion.”

As Ahlberg, Clack, and Blom pointed out, though, their conclusion was preliminary. Additional study and modeling of the tetrapod’s joints would hopefully better constrain our ideas about how Ichthyostega moved. That research has now been published. In today’s issue of Nature, anatomists Stephanie Pierce, Jennifer Clack, and John Hutchinson presented the results of their investigation into how Ichthyostega got around.

Using a high-resolution, three-dimensional computer model made from skeletal scans, Pierce, Clack, and Hutchinson put their reassembled Ichthyostega through its paces. And, just as the 2005 Nature paper suggested, Ichthyostega wasn’t capable of the classic side-to-side walk used by modern newts and salamanders. In fact, the strange anatomy of Ichthyostega constrained the animal’s movements in such a way that the tetrapod must have used an unusual mode of transportation for a creature with limbs.

The tetrapod’s forelimbs were very limited in their range of motion, the researchers found, and the hindlimbs of Ichthyostega would have been virtually useless for support on Devonian mudbanks. “[T]he most likely mode of forelimb movement on land/substrate,” Pierce and co-authors concluded, “involved synchronous mudskipper-like ‘crutching’ motions.” Ichthyostega didn’t so much walk as plop around supported by its arms.

The skeleton of Ichthyostega was a weird mosaic of features relating to both life in the water and occasional forays on land. The forelimbs of Ichthyostega and similar tetrapods were the first to be co-opted for moving on land, with the hips and hindlimbs following later. And, even then, the forelimbs of Ichthyostega were not novel appendages that evolved for the purpose of walking. Crutching around  in the mud was an extra advantage to limbs that evolved to allow “both station-holding and lifting of the head of of the water to breathe and potentially feed.” Strange as it seems, the limbs of Ichthyostega had more to do with life in the water than on land.

And there’s another little wrinkle to this story. Two years ago, paleontologist Grzegorz Niedźwiedzki and co-authors announced that they have found exceptionally early tetrapod tracks – footprints and trackways dating almost 20 million years before Tiktaalik. Tetrapods capable of walking on land might have evolved far earlier than anyone expected.

There was already plenty of reason to approach these trace fossils with a skeptical eye. For one thing, trace fossils created by invertebrates have frequently been mistaken for early tetrapod footprints. But the new study by Pierce, Clack, and Hutchinson provide another reason to take another look at those controversial traces. If other early tetrapods were like Ichthyostega, they would have been physically incapable of making the kinds of alternating trackways Niedźwiedzki and colleagues described. Trackways created by Ichthyostega and similar tetrapods would probably look like sets of parallel hand prints with a body impression in the middle. Pierce and colleagues caution that as-yet-unknown tetrapods with different capabilities might have created the especially old tracks, but the gap in time, resemblance of the tracks to invertebrate traces, and the inability of Ichthyostega-like tetrapods to make the tracks raises the question of what, exactly, the trace fossils from Poland represent.

Evolution defies simplistic storytelling. As much as I loved the story of the Eusthenopteron that could, that narrative has been completely tossed out. Even after specialized limbs evolved, early tetrapods had not yet settled in to life on land. What was once a nice, neat story has transmuted into a wonderful mystery. When did those fishy Devonian tetrapods start to spend more time ashore, and how did the change happen? Ichthyostega, mudskipping along in the ancient Devonian, is a fantastic clue that we are only just getting acquainted with the creatures that passed down the body plan that lets me sit here, typing away with my extremely-modified fins.

For more on Ichthyostega and early tetrapods, see John Hutchinson’s post, Carl Zimmer’s essay, and the chapter “From Fins to Fingers” in my book Written in Stone.


Ahlberg, P., Clack, J., & Blom, H. (2005). The axial skeleton of the Devonian tetrapod Ichthyostega Nature, 437 (7055), 137-140 DOI: 10.1038/nature03893

Niedźwiedzki, G., Szrek, P., Narkiewicz, K., Narkiewicz, M., & Ahlberg, P. (2010). Tetrapod trackways from the early Middle Devonian period of Poland Nature, 463 (7277), 43-48 DOI: 10.1038/nature08623

Pierce, S., Clack, J., & Hutchinson, J. (2012). Three-dimensional limb joint mobility in the early tetrapod Ichthyostega Nature DOI: 10.1038/nature11124

6 thoughts on “A New Twist in the Tetrapod Tale

  1. Hi Brian,

    Grzegorz Niedzwiedki and I are in the process of writing a full description of the Zachelmie trackway material, which I don’t want to pre-empt, but a few things are worth flagging up:

    1) Although we have a single short lateral-sequence-walk track at Zachelmie (Fig. 2a in Niedzwiedzki et al. 2010), the majority of the tracks from the site show one pair of appendages being moved synchronously, or switching back and forth between synchronous and alternating movement. Fig. 2c is a good example of this type of track, as is the cover picture of Nature. It is clear that the tracks were made subaquatically (although possibly in very shallow water), because in addition to the bipedal tracks we also have a number of genuine ‘singleton prints’, i.e. single footprints that are isolated on large sediment surfaces (as opposed to being found on small broken-off blocks). One such print can be seen at the top of Fig 2a (it is shown with vertical hatching in the interpretative diagram) and another with obvious digit marks and a big unilateral displacement rim is visible on the cover picture just to the left of the near end of the trackway. These isolated prints presumably represent swimming animals taking an occasional ‘punt’ against the substrate with a single limb.

    The interesting point here is that the predominant movement pattern at Zachelmie fits perfectly well with an Ichthyostega-like animal pushing itself along in shallow water by synchronous or alternating use of its forelimbs. Of course the lateral-sequence-walk track still has to be explained (a point I will return to below), but the fact of the matter is that the Zachelmie tracks are the LEAST challenging of the putative Devonian tetrapod tracks to interpret with reference to Pierce et al’s Ichthyostega model. The ones that actually cause problems are the Valentia Island tracks from Ireland (latest Middle Devonian, 20 mya older than Ichthyostega) and the Genoa River tracks from Australia (Late Devonian, exact age uncertain but probably more or less contemporary with Ichthyostega). These are very obvious tetrapod walking tracks, in the case of Genoa River with well-preserved digit impressions, and they all show lateral sequence walks. So whether or not we want to include the Zachelmie tracks in the equation, we have to face the fact that tetrapods of some sort or another were performing lateral sequence walks well before the time of Ichthyostega.

    2) It is not a matter of conjecture but of observed fact that Ichthyostega in some ways is a very peculiar and autapomorphic animal. As Pierce et al. note, the limbs are reasonably standard-issue for known stem tetrapods, but the crucial point in the equation is that the vertebral column is not. Ichthyostega’s axial skeleton, including both vertebral column and ribcage, is extremely distinctive and shows adaptations for rigidity that are absent in all the other really primitive tetrapods known to us. This is likely to have a big impact on stride length and the capacity to perform a lateral sequence walk. Pierce et al argue that long-axis rotation of the humerus and femur is an essential prerequisite for this style of walking, but here I disagree. It may indeed be necessary for terrestrial locomotion of this type, where the body is lifted off the ground by the limbs, but aquatic lateral-sequence walking (which appears to be what we see in those Valentia Island and Genoa River tracks that lack body drags) only seems to require a flexible vertebral column and appendages that project laterally from the body. Here’s an epaulette shark:

    So to argue from the Ichthyostega model that Devonian tetrapods in general couldn’t perform lateral sequence walks, and that the contemporary fossil trackways of digited feet performing such walks must therefore have been made by Something Else (unspecified), doesn’t really stand up.

    3) Just for the record, there’s nothing about the Zachelmie tracks that suggests an arthropod identity. No prints of multiple narrow-ended appendages, no telson drags, and no sign of the whole ‘sharp-edged’ quality so typical of prints made by sclerotised cuticle. 

    As I said, a detailed description of all the Zachelmie material is in the pipeline; we hope to submit it sometime during the autumn. This field certainly has moved a long way from Eusthenopteron crawling heroically out of the ooze, and the Pierce et al paper forms a valuable contribution to the emerging picture of what really happened. I dare say you will have occasion to keep blogging about the origin of tetrapods over the next few years.

    Cheers,  Per

    1. I appreciate Per’s detailed comments and look forward to more details in publication(s) on the Zachelmie tracks, which are of course a very important discovery that deserves more detail. We agree, and Clack in particular has co-published on with Per, that *some* details of the anatomy of Ichthyostega seem to be autapomorphic based on evidence currently available from a few basal tetrapods, most of which are quite fragmentary and often hard to compare with Ichthyostega in critical regions. That is a crux of the problem.

      Another, perhaps equally important, point is that vertebral morphology needs to be reconstructed in 3D using similar methods to ours, and then used to examine vertebral flexibility for Ichthyostega as well as other tetrapods and compare them. Until then, the effects of vertebral flexibility on what kinds of locomotion are possible are not well known, which we alluded to in the paper. Per and colleagues are joining us in some of this work.

      But to some degree the limb anatomy of Ichthyostega may not be autapomorphic for the most basal tetrapods, in particular features germane to assessing joint ranges of motion as in our study. And potentially, vertebral mobility in the “lumbar” region (where most lateral undulation should occur) might not be so autapomorphic. We commented on this in the paper, particularly in the Supp Info. Of course, we have precious little data for hindlimb joints of some of these taxa (largely just Acanthostega and Ichthyostega), but these taxa seem to agree in those germane features so it is possible (but not yet unambiguous, as we cautioned) that these features were ancestral along the tetrapod stem. If so, a key question would be when those features then changed, but that is not yet known. If not, then there are alternative explanations including autapomorphy/convergence in Ichthyostega or Acanthostega. More evidence is needed to sort out the most parsimonious answer.

      The long-axis rotation of the hip, however, is pivotal for planting the pes flat on the substrate, because the knee and ankle do not show much capacity for such motions. That is part of the reason we concluded that Ichthyostega, and perhaps some other stem tetrapods, could not walk normally, especially on land (bottom walking/shallow paddling is another issue).

      Another crux of any disagreement (and experts agree on a lot of details, it should be noted) is what taxa known from skeletal remains, or what specific ancestral morphologies along the phylogeny, made known footprints. This is always a bugbear for ichnology; matching body fossils and synapomorphies is standard practice. All good data for limb material comes from Late Devonian taxa, and Per notes the ages of footprints that are known, many of which are 20+ million years older than Ichthyostega/Acanthostega. The trackways do not show what the specific body morphology was like, especially proximal joint mobility. We do not argue that tetrapods could not have made them, but that “early tetrapods with the skeletal morphology and limb mobility of Ichthyostega were unlikely to have made some” of them; particularly symmetrical (inferred lateral sequence) trackways.

      The key unresolved question is, what was body/limb morphology like in stem tetrapods 20+ million years before Ichthyostega/Acanthostega, or in any likely trackmakers? One possible answer is it was identical to these taxa (i.e., plesiomorphic), requiring relative evolutionary stasis during a period of “locomotor experimentation” and apparent great diversification from the Middle Devonian to Late Devonian. This might seem implausible. Another is that those earlier tetrapods (or other ones, in the case of Late Devonian trackways) were quite different and perhaps had greater ranges of joint motion as a result; that might be more plausible. And there are other possible explanations too. We aren’t sure that early tetrapod morphology is so completely known that this question can be answered at present, which is why we erred on the side of caution in our paper and left the issue open. We don’t know how many tetrapod taxa existed in the Middle-Late Devonian and how much locomotor morphology varied among these. Plausibility is not enough for scientists of course; our paper was about basing functional interpretations on the best morphological evidence we could gather, and interpreting that phylogenetically to the degree we could. New evidence, particularly for the precise morphology of the skeletons of Devonian tetrapods, could change any interpretation, as the history of this field has long been teaching experts.

      1. Just to flesh this out a bit further for those who are not familiar with the details, one of the real puzzles of the story is the transformation that the hip and shoulder joints undergo at the fish-tetrapod transition. In fish members of the tetrapod stem group, like Eusthenopteron or Sauripterus, these joints have spheroidal surface curvatures and plainly allowed rotatory movements of the humerus and femur, but they face posteriorly with only a small lateral component. The fins probably had movement ranges broadly similar to what we see in Latimeria or the Australian lungfish Neoceratodus. The picture is less clear in the most tetrapod-like fishes, Panderichthys and Tiktaalik, but the shoulder joint at least seems to retain a rotatory capacity and still faces posteriorly – though perhaps with a slight lateral shift. 

        But when we get to the earliest tetrapods, we find something strange: the joint surfaces of the hip and shoulder now face laterally, but they are elongated and twisted into weird spiral curvatures. Furthermore, the head of the femur, which was previously a straightforward oval-ish ball, now has an obliquely sloping upside-down-horseshoe shape. The hollow of the horseshoe accommodates a ventral buttress that braces the lower edge of the hip socket. At the same time the pelvis has become enormously enlarged compared to the ancestral fish condition, it now attaches to the vertebral column via a sacral rib, and large new areas for muscle attachment have appeared ventral and posterior to the hip socket.

        So it seems that at the fish-tetrapod transition,

        1) the paired appendages are reoriented from posterior to lateral (as shown by the sockets on the girdles);

        2) the appendages take on an enhanced locomotory role, the hind limb in particular becoming much more important as a propulsive organ than the fish pelvic fin (as shown by the size increase and reconfiguration of the pelvis);

        3) but the movement ranges of the limbs are greatly constrained (as shown by the detailed morphology of the joint surfaces).

        It certainly isn’t easy to make sense of this functional package. 

  2. back when i was galavnting around adromeda destrying up to 760 hundered planet till someone moved a star.they all used mirriors and then got turned into animals when theywere destryoed

  3. The other half of the evidence, wholly ignored here, is that zillions of species exactly did die out. The weight of evidence is that, faced with the crisis of environmental change, nothing evolved, nothing adapted, nothing lived thru it except those already fitting multiple environmental niches.
    For some reason, the evolution mythology can’t stomach that fact.

Leave a Reply

Your email address will not be published. Required fields are marked *