News & Views | Published:


Between water and land

The most informative examples of large-scale evolution are provided by major transitions between environments. Fresh research on an ancient amphibian shows how it adapted to locomotion both in water and on land.

One of the defining events in the history of life was the emergence of terrestrial vertebrates from early fish. The oldest known fossils that illustrate the transition to land are those of Ichthyostega from the Upper Devonian of East Greenland. This 360-million-year-old amphibian resembled fish in many features of its skeleton, but possessed pelvic (hip) and pectoral (shoulder) girdles and limbs capable of supporting the body and allowing movement on land (Fig. 1). Ichthyostega was initially described by Säve-Söderbergh1 and Jarvik2,3, but many aspects of its anatomy remained unknown, limiting understanding of aspects of the transition from water to land.

Figure 1: Reconstructions of the body form of Ichthyostega, one of the oldest known amphibians.

Length of the animals is about 1 m. These depictions show their capacity for swimming in water using the tail, or walking on land with the trunk supported by muscles attached to specialized neural spines on the vertebrae. (Reconstruction by Tony Terenzi, based on ref. 4.)

On page 137 of this issue, Ahlberg, Clack and Blom4 provide a new reconstruction and functional analysis based on the original specimens, as well as on many fossils collected subsequently by Clack and her colleagues. Their work reveals numerous specializations in the components of the main body axis — the axial skeleton — that were not previously recognized. They demonstrate that Ichthyostega possessed a unique vertebral structure that enabled the trunk to be supported above the ground, and that it also had an elaboration of the ribs at the base of the tail that facilitated swimming.

Unlike other primitive amphibians and their plausible antecedents among lobe-finned fish, the configuration of the trunk vertebrae in Ichthyostega is not uniform along the column, but instead is regionally specialized. Much of the interpretation hinges on variation in the vertebrae, each of which consists of a crescentic ventral portion (the centrum) partially surrounding the notochord, and a neural arch that encloses the spinal cord and extends dorsally as a neural spine.

In Ichthyostega, the neural spines of the vertebrae behind the shoulder girdle are angled backwards, while those above and in front of the pelvic girdle are angled forwards. Those in the middle of the trunk are more nearly vertical, but have pronounced areas of roughness, as occur in modern vertebrae for the attachment of muscles and associated tendons. Because muscles produce their maximum force when acting at right angles to the structures to which they are attached, the varied geometry of the neural spines indicates that the attached muscles could have raised the central portion of the trunk, lifting it above the ground. This configuration is analogous to that of a suspension bridge, with the trunk lifted by muscles (acting like cables) running from elevated supports (the pelvic and pectoral girdles and limbs).

Ahlberg et al.4 point out that the widely expanded anterior ribs, also unique to Ichthyostega, would have made the spinal column more rigid, but would also have greatly restricted lateral undulation of the trunk, which is a major factor in aquatic locomotion. The great length of these ribs, compared with their size in lobe-finned fish, would also have prevented collapse of the chest cavity and lungs as Ichthyostega dragged itself out of the water. The rigidity of the anterior trunk appears to have been compensated for by enlargement of the ribs at the base of the tail, which would have increased the area for attachment of muscles that moved the tail from side to side.

The lineage that included Ichthyostega thus seems to have undergone the transition from water to land through changes that would have increased the facility for movement in both environments. What remains to be learned of the skeleton of Ichthyostega includes the anatomy of the wrist and hand, and the orientation of the rear limb when walking on land.

Surprisingly, the particular design observed in Ichthyostega was ultimately unsuccessful, for there are few if any fossils known after about 360 million years ago that represent plausible descendants of this lineage. Rather than elaboration of the neural spines and their musculature to support the vertebral column, the more successful terrestrial vertebrates expanded the vertebral centra, the same elements that provide the major support in all modern terrestrial vertebrates. Homologues of the centra in human vertebrae were little, if at all, ossified in Ichthyostega. It is interesting to note that the development of the different elements of the vertebrae is under different genetic control. The dorsal and lateral portions of the neural arch are controlled by bone morphogenetic protein 4 (BMP-4) and the vertebral centra by Pax1 (ref. 5). Random mutations in one or the other of these developmental pathways could have led to alternative directions of evolutionary change.

From the standpoint of patterns of evolution, there are at least 11 lineages of advanced lobe-finned fish and early amphibians from the Upper Devonian6, only one of which is a plausible close relative of later land vertebrates. What one sees from the fossil record is a 15-million-year history of ‘experimentation’ among the descendants of fish that had already evolved fins with a central bony axis, a swim bladder (the fish equivalent of lungs) and paired nostrils opening into the mouth7. Comparable diversity of ultimately unsuccessful experiments is also evident in other major transformations in the evolution of vertebrates, including the early evolution of mammals8, the origin of whales9 and evolution within our own genus Homo10. In the last case, as many as ten recognized species have diverged within the past 2 million years, differing greatly in stature, cranial capacity and facial structure, only one of which remains to give rise to future generations.


  1. 1

    Säve-Söderbergh, G. Medd. Grønland 98, 1–211 (1932).

  2. 2

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

  3. 3

    Jarvik, E. Fossils Strata 40, 1–206 (1996).

  4. 4

    Ahlberg, P. E., Clack, J. A. & Blom, H. Nature 437, 137–140 (2005).

  5. 5

    Christ, B., Huang, R. & Scaal, M. Anat. Embryol. 208, 333–351 (2004).

  6. 6

    Clack, J. Amphibian Biology Vol. 4. 979–1029 (Surrey Beatty, Chipping Norton, NSW, Australia, 2000).

  7. 7

    Zhu, M. & Ahlberg, P. Nature 432, 94–96 (2004).

  8. 8

    Luo, Z. -X., Kielan-Jaworowska, Z. & Cifelli, R. Acta Palaeontol. Pol. 47, 1–78 (2002).

  9. 9

    Thewissen, J. & Williams, E. M. Annu. Rev. Ecol. Syst. 33, 73–90 (2002).

  10. 10

    Lahr, M. & Foley, R. Nature 431, 1043–1044 (2004).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Further reading

  • Astrophysics in 2005

    • Virginia Trimble
    • , Markus J. Aschwanden
    •  & Carl J. Hansen

    Publications of the Astronomical Society of the Pacific (2006)


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.