248 MK.LION Years Ago 



Scutosaurus 



Intermediate Pareiasaur 



Ten feet long 



Fused bony plates over shoulder and pelvic 



areas. Unfused plates spread laterally 



An 



248 Million Years Ago 



Anthodon 



Advanced Pareiasaur 



Three feet long 



interlocking mosaic of plates 



covered its back and sides 



210 Million Years Ago 



Proganochelys 



Primitive Turtle 



Three feet long 



Looked like a modem turtle; shell 



completely enveloped its body 



<: 



o 



o 

 :3 



shoulder to the shell. Again, this structure 

 first arose in early pareiasaurs and must 

 have initially served a different function — 

 one that is clarified by comparison with 

 therapsids (primitive ancestors of mam- 

 mals). Therapsids, which are totally unre- 

 lated to pareiasaurs and turtles, had inde- 

 pendently evolved a similar structure, and 

 research suggests that the acromion proc- 

 ess improved the flexibihty of the shoulder 

 region. In primitive reptiles, the shoulder 

 blade and collarbone are rigidly cormected 

 to each other along their entire length. In 

 therapsids, the shoulder blade meets the 

 collarbone only at the acromion process, 

 and the two bones can move with respect 

 to each other. The acromion process un- 

 doubtedly served a similar function in 

 pareiasaurs. Thus, the acromion process 

 initially evolved in pareiasaurs as a mobile 

 articulation between the shoulder bones, 

 and initially had nothing to do with an- 



choring the shoulder blade to the shell. 

 Not surprisingly, in the most primitive tur- 

 tie, Proganochelys, the acromion process 

 retains the old function, and meets the col- 

 larbone, not the shell. Only in more ad- 

 vanced turtles did it shift position and 

 come into contact with the shell. 



The highly distinctive body plan of tur- 

 tles, therefore, did not arise in one huge 

 evolutionary leap. Rather, traits that were 

 evolutionary holdovers from their 

 pareiasaur ancestors were modified and in- 

 tegrated with one another. The shell 

 started out as merely a supporting row of 

 bony plates in pareiasaurs. Later, in turtles, 

 these were co-opted to form the basis of a 

 rigid protective, insulated box. Similarly, 

 the short stubby body, the acromion proc- 

 ess, and many other turtle traits (such as 

 the tall, columnlike shoulder blade and 

 oddly shaped arm and thigh bones) are 

 often thought to have arisen purely as 



adaptations to life in a shell. Yet they ap- 

 peared first in pareiasaurs, long before the 

 shell appeared. 



All this illustrates once again the 

 serendipity of evolution. Natural selection 

 favors what's best now — which is rarely 

 what's best in the long run. Usually, this 

 means adapting an existing organ to per- 

 form some new role tolerably well, instead 

 of going back to the drawing board and 

 evolving a completely new structure that 

 does the job perfectly. (Vertebrate history 

 is full of such makeshift expediency: our 

 arms are really only modified forelegs, and 

 our ear bones arose from bits of gill and 

 jaw.) Traits that originally served other 

 purposes in their pareiasaur forebears be- 

 came modified in turtles to serve functions 

 related to life in a shell. And so well have 

 they been integrated into their new roles 

 that it is difficult for us to imagine them as 

 having evolved to do anything else. 



Anthodon, a 



pareiasaur, also had 



five vertebrae in its 



neck, but only fourteen 



in its back. In front of 



the wide, barrel-shaped 



rib cage, there was a narrow 



shoulder girdle. Shifting the girdle three 



vertebrae farther back into the rib cage 



would result in a turtlelike arrangement. 



Shoulder blade 



Proganochelys, the most primitive turtle 



known, had eight neck vertebrae and ten 



back vertebrae, and the shoulder 



girdle lay within the rib cage. 



The ribs are fused with 



the shell. 



Acromion process 



65 



