leading edge of die foot, trailed by four stout ones. 

 Instead of a weight-bearing foot with five toes, as 

 pictured by Jarvik, we found that the entire hind leg 

 was shaped like a paddle [see photograph on page 41]. 

 An analogous appendage, which evolved indepen- 

 dently (and with five toes, not seven), is the fore- 

 limb of the modern river dolphin. 



That discovery alerted us to other problematic as- 

 pects of Ichthyostega's anatomy, as it had originally been 

 conceived. To make accurate estimates of the relative 

 sizes ofbody parts, paleontologists ideally need a head, 

 a trunk, a forelimb, and a foot from the same indi- 

 vidual. We lacked such a complete package; our best 

 evidence showed that instead of having hind limbs 

 larger than forelimbs, as in most tetrapods, Ichthyoste- 

 ga had big, strong forelimbs paired with diminutive 

 hind limbs. Was that real- 

 ly the case, or was this con- 

 clusion merely an artifact 

 of examining a composite 

 skeleton? 



To investigate further 

 the apparent anomaly of 

 leg size, I organized a sec- 

 ond expedition to Green- 

 land in 1998, aiming to 

 find more Iclithyostega 

 material. Quite by chance, 

 all four of our team mem- 

 bers were female, so we 

 called ourselves the Girls 

 in Greenland expedition. 

 Again, we were fortunate 

 in having the resources of 

 the Danish geologists to 

 help us. As so often hap- 

 pens in collecting expedi- 

 tions, we discovered our 

 best material in the last few 



days of the five-week field season, but it proved cru- 

 cial. We collected shoulder, forelimb, trunk, and hind- 

 limb bones, all from the same individual animal, which 

 gave us the correct proportions for the limbs. 



We also found a well-preserved skull from a dif- 

 ferent individual. Gompared with the rest of the 

 skeleton, skulls are complex and full of information. 

 They not only enable paleontologists to determine 

 relations among groups of animals, but in the case 

 of early tetrapods, they also reflect changes that were 

 as vital to life on land as limbs were. For example, 

 Iclithyostega skulls indicated how breathing organs 

 and ear regions derived from the gill areas of fish. 



Drawing together our fresh material and the pre- 

 viously collected specimens in the Cambridge lab, 

 we examined all our fossils with newer mechanical 



techniques and also with the newly developed proc- 

 ess of CT scanning, specially geared to see through 

 rock. We gave Iclithyostega a complete makeover. 

 With most tetrapods, including Acatithostega, we 

 could be reasonably sure which parts — ears, for in- 

 stance — corresponded from one animal to anoth- 

 er. But Iclithyostega was so weird in the regions at 

 the back of the head, where the ears and brain had 

 been, that interpreting their evolution had never 

 been possible before. CT imaging and modeling 

 helped us understand the anatomy of that area. We 

 were able to conclude that the animal had a high- 

 ly specialized ear that was probably adapted for use 

 in water. For example, it might have enabled 

 Iclithyostega to listen for aquatic prey. Alternatively — 

 if we infer that Iclithyostega was itself able to pro- 



inches 



Recent skeletal reconstruction of Ichthyostega (top) shows the unusual shape 

 of the back, the paddlelike hind limbs, and the broad overlapping ribs. No 

 "hands" of the creature have been found, but the "arms" and shoulders 

 were strong and could hold up the front part of the body. A skeletal recon- 

 struction of its contemporary, the more aquatic Acanthostega (above), shows 

 a body better adapted for swimming. All four limbs were paddlelike, and the 

 tail fin was deeper and longer than it was in Ichthyostega. 



duce sound to communicate — the ear might have 

 helped individuals hear one another and perhaps 

 find mates. 



The next part of Ichthyostega we reconsidered was 

 the vertebral column, or backbone. Earlier re- 

 constructions had shown it as made up of elements 

 that were more or less uniform along the entire length 

 of the backbone, as is the case in Acanthostega. The 

 neural spines (equivalent to the bumps you can feel 

 when you run your finger up the middle of your back) 

 were all shown pointing backward and all roughly the 

 same length. Instead, we noticed that the neural spines 

 changed their direction progressively from the shoul- 

 der to the hip [see upper illustration above]. In the cen- 

 tral, thoracic region the spines pointed conventional- 



July/August 2006 NATURA1 HISTORY 



3^ 



