News & Views | Published:


A firm step from water to land

A project designed to discover fossils that illuminate the transition between fishes and land vertebrates has delivered the goods. At a stroke, our picture of that transition is greatly improved.

The concept of ‘missing links’ has a powerful grasp on the imagination: the rare transitional fossils that apparently capture the origins of major groups of organisms are uniquely evocative. But the concept has become freighted with unfounded notions of evolutionary ‘progress’ and with a mistaken emphasis on the single intermediate fossil as the key to understanding evolutionary transitions. Much of the importance of transitional fossils actually lies in how they resemble and differ from their nearest neighbours in the phylogenetic tree, and in the picture of change that emerges from this pattern.

We raise these points because on pages 757 and 764 of this issue1,2 are reports of just such an intermediate: Tiktaalik roseae, a link between fishes and land vertebrates that might in time become as much of an evolutionary icon as the proto-bird Archaeopteryx. Several specimens have been found in Late Devonian river sediments on Ellesmere Island in Nunavut, Arctic Canada. They show a flattened, superficially crocodile-like animal, with a skull some 20 centimetres in length. The body is covered in rhombic bony scales, and the pectoral fins are almost-but-not-quite forelimbs; these contain robust internal skeletons, but are fringed with fin rays rather than digits. Tiktaalik goes a long way — but not quite the whole way — towards filling a major gap in the picture of the vertebrate transition from water to land.

It has long been clear that limbed vertebrates (tetrapods) evolved from osteolepiform lobe-finned fishes3, but until recently the morphological gap between the two groups remained frustratingly wide. The gap was bounded at the top by primitive Devonian tetrapods such as Ichthyostega and Acanthostega from Greenland, and at the bottom by Panderichthys, a tetrapod-like predatory fish from the latest Middle Devonian of Latvia (Fig. 1). Ichthyostega4 and Acanthostega5 retain true fish tails with fin rays but are nevertheless unambiguous tetrapods with limbs that bear digits6. Panderichthys7 is vaguely crocodile-shaped and, unlike the rather conventional osteolepiform fishes farther down the tree, looks like a fish–tetrapod transitional form. The shape of the pectoral fin skeleton and shoulder girdle are intermediate between those of osteolepiforms and tetrapods, suggesting that Panderichthys was beginning to ‘walk’, but perhaps in shallow water rather than on land8.

Figure 1: Tiktaalik in context.

The lineage leading to modern tetrapods includes several fossil animals that form a morphological bridge between fishes and tetrapods. Five of the most completely known are the osteolepiform Eusthenopteron16; the transitional forms Panderichthys17 and Tiktaalik1; and the primitive tetrapods Acanthostega and Ichthyostega. The vertebral column of Panderichthys is poorly known and not shown. The skull roofs (left) show the loss of the gill cover (blue), reduction in size of the postparietal bones (green) and gradual reshaping of the skull. The transitional zone (red) bounded by Panderichthys and Tiktaalik can now be characterized in detail. These drawings are not to scale, but all animals are between 75 cm and 1.5 m in length. They are all Middle–Late Devonian in age, ranging from 385 million years (Panderichthys) to 365 million years (Acanthostega, Ichthyostega). The Devonian–Carboniferous boundary is dated to 359 million years ago18.

Panderichthys lived about 385 million years ago at the end of the Middle Devonian; Ichthyostega and Acanthostega lived about 365 million years ago during the Late Devonian. However, the earliest fragmentary tetrapods from Scotland9,10 and Latvia9 date back to perhaps 376 million years ago, so the morphological gap between fish and tetrapod corresponds to a time gap of under 10 million years.

Into this gap drops Tiktaalik. The fossils are earliest Late Devonian in age, making them at most 2 million or 3 million years younger than Panderichthys. With its crocodile-shaped skull, and paired fins with fin rays but strong internal limb skeletons, Tiktaalik also resembles Panderichthys quite closely. The closest match, however, is not to Panderichthys but to another animal, Elpistostege, from the early Late Devonian of Canada. Elpistostege is known only from two partial skulls and a length of backbone, but it has long been recognized as a fish– tetrapod intermediate11,12, probably closer to tetrapods than is Panderichthys. This impression is now confirmed: the authors1,2 demonstrate convincingly that Elpistostege and Tiktaalik fall between Panderichthys and the earliest tetrapods on the phylogenetic tree.

So, if Tiktaalik is in effect a better-preserved version of Elpistostege, why is it important? First, it demonstrates the predictive capacity of palaeontology. The Nunavut field project had the express aim of finding an intermediate between Panderichthys and tetrapods, by searching in sediments from the most probable environment (rivers) and time (early Late Devonian). Second, Tiktaalik adds enormously to our understanding of the fish–tetrapod transition because of its position on the tree and the combination of characters it displays.

In some respects, Tiktaalik and Panderichthys are straightforward fishes: they have small pelvic fins13, retain fin rays in their paired appendages and have well-developed gill arches, suggesting that both animals remained mostly aquatic. In other regards, Tiktaalik is more tetrapod-like than Panderichthys. The bony gill cover has disappeared, and the skull has a longer snout (Fig. 1). These changes probably relate to breathing and feeding, which are linked in fishes because the movements used for gill ventilation can also be used to suck food into the mouth. A longer snout suggests a shift from sucking towards snapping up prey, whereas the loss of the gill cover bones (which turned the gill cover into a soft flap) probably correlates with reduced water flow through the gill chamber. The ribs also seem to be larger in Tiktaalik, which may mean it was better able to support its body out of water1. The only real peculiarity of Tiktaalik is its poorly ossified vertebral column that seems to contain an unusually large number of vertebrae.

These character distributions paint an intriguing picture. Tiktaalik is clearly a transitional form, more tetrapod-like than Panderichthys in its breathing and feeding apparatus, but with similar locomotory adaptations. Crucially, because Tiktaalik occupies a position closer to tetrapods on the tree than does Panderichthys, their shared characters can be inferred to be attributes of the segment of the tree between the branches that carry the two animals (Fig. 1, red). Panderichthys showed us a morphology that could be interpreted as directly intermediate between osteolepiform and tetrapod. But only the similar yet ‘upgraded’ morphology in Tiktaalik demonstrates that this interpretation is correct: this really is what our ancestors looked like when they began to leave the water.

Two aspects of Tiktaalik's anatomy relate to the origin of new structures in tetrapods: the ears and limbs. The tetrapod middle ear has arisen as a modification of the fish spiracle (a small gill slit) and hyomandibula (a bone supporting the gill cover). Panderichthys possesses a widened spiracle, interpreted as the intake for air or water, and a shortened hyomandibula14. Tiktaalik shows an almost identical condition, but with an even wider spiracle, indicating that this morphology too is genuinely transitional.

The pectoral fin skeleton of Tiktaalik is notable not only because of its transitional nature, but also because its excellent preservation has allowed the individual bones to be freed of the rock and manipulated to estimate ranges of movement2. It turns out that the distal part of the skeleton is adapted for flexing gently upwards — just as it would if the fin were being used to prop the animal up. Although these small distal bones bear some resemblance to tetrapod digits in terms of their function and range of movement, they are still very much components of a fin. There remains a large morphological gap between them and digits as seen in, for example, Acanthostega: if the digits evolved from these distal bones, the process must have involved considerable developmental repatterning. The implication is that function changed in advance of morphology.

The body form represented by Tiktaalik and Panderichthys was evidently an actual step on the way from water to land. Just over 380 million years ago, it seems, our remote ancestors were large, flattish, predatory fishes, with crocodile-like heads and strong limb-like pectoral fins that enabled them to haul themselves out of the water. Further information will emerge from the full description of the fossils, and from detailed comparisons with Devonian tetrapods such as the very primitive Ventastega15.

Of course, there are still major gaps in the fossil record. In particular we have almost no information about the step between Tiktaalik and the earliest tetrapods, when the anatomy underwent the most drastic changes, or about what happened in the following Early Carboniferous period, after the end of the Devonian, when tetrapods became fully terrestrial. But there are still large areas of unexplored Late Devonian and Early Carboniferous deposits in the world — the discovery of Tiktaalik gives hope of equally ground-breaking finds to come.


  1. 1

    Daeschler, E. B., Shubin, N. H. & Jenkins, F. A. Jr, Nature 440, 757–763 (2006).

  2. 2

    Shubin, N. H., Daeschler, E. B. & Jenkins, F. A. Jr Nature 440, 764–771 (2006).

  3. 3

    Cope, E. D. Proc. Am. Phil. Soc. 30, 278–281 (1892).

  4. 4

    Ahlberg, P. E. et al. Nature 437, 137–140 (2005).

  5. 5

    Coates, M. I. Trans. R. Soc. Edinb. Earth Sci. 87, 363–421 (1996).

  6. 6

    Clack, J. A. Gaining Ground: The Origin and Early Evolution of Tetrapods (Indiana Univ. Press, Bloomington, 2002).

  7. 7

    Vorobyeva, E. I. & Schultze, H. -P. in Origins of the Higher Groups of Tetrapods (eds Schultze, H.-P. & Trueb, L.) 68–109 (Comstock, Ithaca, 1991).

  8. 8

    Vorobyeva, E. I. & Kuznetzov, A. in Fossil Fishes as Living Animals (ed. Mark-Kurik, E.) 131–140 (Inst. Geol., Tallinn, Estonia, 1992).

  9. 9

    Ahlberg, P. E. Nature 373, 420–425 (1995).

  10. 10

    Ahlberg, P. E. Zool. J. Linn. Soc. 122, 99–141 (1998).

  11. 11

    Westoll, T. S. Nature 141, 127–128 (1938).

  12. 12

    Schultze, H. -P. & Arsenault, M. Palaeontology 28, 293–309 (1985).

  13. 13

    Boisvert, C. Nature 438, 1145–1147 (2005).

  14. 14

    Brazeau, M. & Ahlberg, P. E. Nature 439, 318–321 (2006).

  15. 15

    Ahlberg, P. E., Lukševičs, E. & Lebedev, O. A. Phil. Trans. R. Soc. Lond. B 343, 303–328 (1994).

  16. 16

    Jarvik, E. Basic Structure and Evolution of Vertebrates (Academic, New York, 1980).

  17. 17

    Ahlberg, P. E. & Milner, A. R. Nature 368, 507–514 (1994).

  18. 18

    Gradstein, F. M. et al. A Geologic Time Scale 2004 (Cambridge Univ. Press, 2005).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.