RVC Website: | Home | Courses | Higher Degrees | Research | Clinical Services | RVC Enterprise | About Us | Contacts | Search |

SML  >  Research Projects

Assessing the Limb Mobility of the Early Four-Legged Vertebrate Animal Ichthyostega

Video 1: Movie of our 3D model of Ichthyostega

Video 2: Movie of maximal forelimb ranges of motion in Ichthyostega

Videos may play best in Internet Explorer and with Flash plugin.

See Images and Videos page for more!

Could it have moved like a salamander, or something else?

Ichthyostega flesh reconstruction by Julia MolnarFlesh reconstruction of the whole body of Ichthyostega by Julia Molnar (jmolnar@rvc.ac.uk). Please credit her for this work if it is used.

Our team of three scientists consists of Dr. Stephanie Pierce (RVC and Cambridge), Professor Jenny Clack (Cambridge), and Professor John Hutchinson (RVC). We have been funded by the NERC for the past three years (2009-2012) to construct the first 3D computer models of the earliest bony animals with four limbs (tetrapod vertebrates) that potentially moved on land. On 23 May, 2012 we published a paper online in Nature in which we present results for the iconic tetrapod called Ichthyostega, which has featured prominently in studies of how vertebrates conquered land for over 50 years. This animal is represented by over 300 isolated specimens (see Images and Videos), many of which we have studied, enabling us to scale and combine them (predominately using the single best-preserved skeleton, which is affectionately known as 'Mr. Magic') into a 3D model, shown below.

3D Model of Ichthyostega

Limb movements

We took microcomputer tomography or µCT scan reconstructions of each bone in the skeleton of Ichthyostega and articulated them in 3D using commercial software (3D Studio Max and SIMM). By manipulating the 3D model we were able to test how the limbs of Ichthyostega could, or could not, have moved in life. Specifically, we estimated the range of motion (links here and here) in each of the major limb joints (shoulder, elbow, hip, knee). The wrist/ankle and fingers/toes were not dealt with in any numerical detail because they have less of an impact on how the whole limb moves, are not preserved (e.g. forefoot or hand) or seem to have had little motion (ankle). Our estimates were made by moving the joints in controlled planes and visually inspecting each limb pose that was allowed or prohibited by the joints.

Extant Animals Modelled in Our StudyValidation

To examine how reliable our model was in predicting limb movement, we built similar skeleton-only models of five modern animals with at least vague resemblances to Ichthyostega, including a crocodile and seal (pictured right). We then compared the skeleton-only ranges of motion to those in cadaveric (dead) animal limbs to ensure our measurements were realistic. This test showed that our computer models always tended to overestimate the ranges of motion, indicating that soft tissues, such as restraining ligaments and bulky muscles, would only decrease the mobility of the joints. As such, our 3D modelling technique tended to over-estimate the range of motion of joints, which ended up being a bias that fortunately strengthened our conclusions.


Our analysis found that Ichthyostega, as compared to modern land animals, did not have much mobility in its shoulder and hip joints, especially for rotating the limbs around their long axis. (Try rotating your whole arm from palm-side up to palm-side down, this is long-axis rotation) Such rotation is important in the movements of essentially all land animals today, and particularly so in animals with more primitive limb motions such as salamanders, lizards and crocodiles. Because we learned from our validation study that the addition of soft tissue would have restricted movement even further, thus relying on bones alone over-estimated joint ranges of motion, we could strongly infer that Ichthyostega could not have moved in ways very similar to living land animals. In addition, we found that the hip anatomy in Ichthyostega would have prevented the hind feet from making firm contact with the ground - meaning it could not have supported its body weight on all four legs or moved like a salamander, as it had previously been reconstructed for many decades.

Locomotion behaviour

Mudskippers (from Wikipedia)Detailed comparison of limb mobility in Ichthyostega with those of modern animals found that its limb movement - and locomotion behaviour - was most similar to that of a mudskipper fish. Mudskippers (pictured right) are decent swimmers but also can move about temporarily on land (video) using synchronous forelimb “crutching” motions - supporting the body on a sort of tripod with the two mobile fins at the front and the tail plus hind fins in back. The limbs of Ichthyostega probably would have behaved in a similar way, i.e. the forelimbs would have pushed the body up and forward while the hindlimbs and tail trailed behind. Perhaps this type of locomotor behaviour (and restricted limb mobility) in the earliest limbed animals helped to stabilise the body during their first tentative 'steps' onto land.


These possible motions give us a search image for the kinds of tracks that early tetrapods - with a similar skeletal anatomy to Ichthyostega - might have made. Indeed, fossilised tracks with at least some of these features, made either on land or in shallow water, are known from Ireland, Poland and other areas. However, some other tracks known from the same areas do not match these predicted footprints—so what animals made these? We cannot say, but at least now the fossil skeleton and footprint record have some degree of concordance that future studies are likely to improve.

What's next?

We are now working on investigating how the vertebral column (spine) of Ichthyostega and its relatives was constructed and how it might have moved, and what this means for how early vertebrates first locomoted on land and how walking (and running?) may have first evolved. Stay tuned!


We thank the NERC for funding!

NERC          UMZC           Cambridge Dept Zoology


Our team

The three authors of the paper
Left to Right: John Hutchinson, Stephanie Pierce and Jenny Clack

Learn more!

  • See images and videos from this research
  • Read the paper: Pierce, S.E., Clack, J.A., Hutchinson, J.R. 2012. Three-dimensional limb joint mobility in the early tetrapod Ichthyostega. Nature, published online.
  • RVC Press Release (.doc file)
  • Previous story on footprints of very early tetrapods; example
  • Previous story on early tetrapod forelimb fossil material; example
  • The pre-tetrapod fish Tiktaalik
  • Long-axis joint rotation: an example from birds (multimedia)
  • Common mistakes/misconceptions about our study:
    • PRONOUNCIATION: Ick-thee-oh-stay-guh (American) or Ick-thee-oh-stee-guh (European)
    • Please do not use the misleading term "missing link."
    • Ichthyostega was not an ancestor of living land animals, but was a distant cousin. It and its similar relatives such as Acanthostega show us what our first ancestors that were moving on land were probably like.
    • These animals were NOT dinosaurs, reptiles, or amphibians. They were called early tetrapods (four-limbed vertebrates); all bony, originally four-footed animals today are members of the tetrapod group whether they have lost their limbs or not.
    • These early tetrapods seem to have mostly lived in the water. Our study supports this notion and shows that animals like Ichthyostega would have been awkward on land, mostly pulling themselves along with their forelimbs. 'Normal' walking patterns like we see today in salamanders should have evolved later; we are not sure exactly when.
    • This study tells us nothing about how earlier or later tetrapods moved or why any of these animals went extinct. It is only specifically about the earliest tetrapods, for which we have a decent fossil record.

Other experts to contact (not a comprehensive list):

On early tetrapods:

On the evolution of locomotion:

Designed and developed by the RVC Electronic Media Unit