o 



o 

 > 



pareiasaurian forebears. These evolution- 

 ary holdovers — clues to the ancestry of 

 turtles — were later lost in more advanced 

 turtles. Nevertheless, a forty-million-year 

 gap, spanning almost the entire Triassic, 

 still exists between the last pareiasaurs and 

 the earliest-known turtles. When turtles 

 first appear in the fossil record, in the late 

 Triassic, they are represented by at least 

 four distinct lineages, suggesting that the 

 group evolved and radiated slightly earlier. 



A pareiasaurian ancestry helps explain 

 how and why the bizarre turtle body plan 

 evolved. The turtle shell is an adaptive 

 marvel. It forms an organic strongbox, 

 into which the extremities can be retracted 

 out of harm's way. It also supports the tur- 

 tle, whose backbone is fused to the rigid 

 shell and whose shoulder girdle is an- 

 chored to the shell by ligaments (other an- 

 imals need muscles to keep these elements 

 in place). Finally, the shell forms a thick, 

 insulating layer, which confers thermoreg- 

 ulatory advantages. Compared with other 

 reptiles of the same size, turtles overheat 

 more slowly on hot days and cool down 

 more gradually on cold nights. 



But which of these demands favored the 

 evolution of the shell? Pareiasaurs supply 

 the answer. Early pareiasaurs possessed a 

 row of bony plates above their back- 

 bone — the first hint of a shell. Recent 

 work by Dino Frey, a German morpholo- 

 gist, suggests that these plates helped the 

 pareiasaur stop its backbone from sagging. 

 Thus, the precursor of the turtle shell that 

 evolved in the large, heavy pareiasaurs, 

 initially served a supporting function. 

 Only in later pareiasaurs and turtles did 

 these plates spread out over the body and 

 provide protection and insulation. All the 

 earliest turtles were found in terrestrial de- 

 posits alongside dinosaurs and possessed 

 stout legs adapted for walking, not swim- 

 ming. So it seems safe to say that they, like 



260 Million Years Ago 



Captorhinus 



Primitive Reptile 



Two feet long 



No bony plates 



pareiasaurs, were land animals. For a long 

 time people had assumed that turtles must 

 have evolved in the water, because of sup- 

 port problems created by the heavy shell. 

 Yet, not only did turtles evolve on dry 

 land, but initially the shell probably served 

 for support. 



Many other distinctive anatomical traits 

 of turtles appear to be, in one way or an- 

 other, adaptations for life in the shell. A 

 straightforward deduction might be that 

 these traits evolved at the same time as the 

 shell, or immediately afterward, and 

 served their function right from the very 

 beginning. Surprisingly, this isn't the case. 

 Consider the turtle's stout body, for ex- 

 ample. It appears to be adapted to fit into 

 the shell. The wide, short shell of turtles is 

 difficult for predators to overturn or get 

 their jaws around. Furthermore, a shell of 

 this shape is easier to maneuver than a 

 long, skinny one (imagine how much diffi- 



255 Million Years Ago 



Bradysaurus 



Primitive Pareiasaur 



Ten feet long 



A row of small, unfused 



bony plates 



culty a Uzard encased in a rigid tube would 

 have getting around). But this body shape 

 actually arose in the pareiasaur ancestors 

 of turtles long before the shell appeared. 

 The earliest-known pareiasaurs lived in 

 southern Africa, which at the time had 

 only just drifted northward out of the 

 Antarctic circle. The climate then was 

 cool. Because short, fat animals lose heat 

 less rapidly than long, thin ones (which is 

 why many animals, ourselves included, 

 curl into a ball when cold), the stout bodies 

 of pareiasaurs probably helped them con- 

 serve precious body heat. Thus the short 

 body of turtles first served a thermoregula- 

 tory function and initially had nothing to 

 do with life in a shell. The body dictated 

 the shape of the evolving shell, not the re- 

 verse. 



Another example concerns a bony proc- 

 ess on the turtle shoulder blade, the acro- 

 mion process, which helps connect the 



Shoulder blade 



The elongated, lizardlike skeleton o/Captorhinus is 

 typical of primitive reptiles. It had five neck 

 vertebrae, twenty back vertebrae, and a shoulder 

 girdle lying outside the rib cage. 



Illustrations by Michael Lee 



Acromion process 



64 Natural History 6/94 



