The Turtle's 



Long-Lost 



Relatives 



Its ancestors evolved many 



turtlelike traits before they 



acquired shells 



by Michael Lee 



A prominent zoologist once quipped 

 that the only thing turtles have done since 

 the Triassic, some 200 million years ago, 

 is survive. This assessment seems a bit 

 harsh, however, considering the variety of 

 environments that they have conquered. 

 Today, mrtles thrive in oceans, rain forests, 

 and deserts. Yet underlying this ecological 

 diversity is a surprising degree of anatom- 

 ical uniformity. No one has any difficulty 

 recognizing a turtle; in all turtles the body 

 is encased within a rigid, bony box. No 

 other animal has a body architecture that is 

 even remotely similar. 



But where do turtles come from? The 

 oldest fossil turtles (along with the earliest 

 dinosaurs) appear abruptly in Triassic 

 rocks, fully developed and without any 

 obvious precursors. Details of their skull 

 suggested that they evolved from a group 

 of primitive reptiles, but none could be 

 readily identified as turtle ancestors. De- 

 spite more than a century of research, the 

 origin of turtles remains a major enigma. 



Such "morphological gaps" are invari- 

 ably seized upon by creationists as evi- 

 dence against evolution. Scientists, aware 

 of the vagaries of the fossil record, attach 

 little importance to such negative evi- 

 dence — the transitional forms may have 

 once existed, but simply have not yet been 

 discovered. Mere ignorance of something 

 does not demonstrate its nonexistence. 

 One is reminded of the crack about the 

 atheist who couldn't prove that God didn't 

 exist — and so took it on faith. Indeed, re- 

 cent paleontological finds have plugged 

 some of the most embarrassing and persis- 

 tent gaps in the continuum of life: Acan- 

 thostega from Greenland, transitional be- 

 tween fishes and amphibians; Ambii- 

 locetus from Pakistan, a seallike link be- 

 tween whales and their terrestrial ances- 

 tors; and Eoraptor, the most primitive di- 

 nosaur yet discovered. 



Stunning finds by intrepid field parties 

 collecting in exotic locations aren't the 

 only way such "missing links" become 

 known to science. Sometimes a careful 

 reappraisal of known forms can yield 

 major surprises. For more than a century, 

 the origin of birds remained a matter of 

 conjecUire, until John Ostrom, a Yale pale- 

 ontologist, pointed out that certain small 

 bipedal, carnivorous dinosaurs closely re- 

 sembled Archaeoptetyx, the first bird — so 

 closely that a specimen of Archaeopteiyx 

 had been mistakenly cataloged as a di- 

 nosaur by museum workers. In a similar 

 vein, I think that I have identified a group 

 of animals, the pareiasaurs, that bridge the 

 huge morphological gap separating the 

 oldest turtles from primitive, lizardlike 

 reptiles. 



Pareiasaurs have been known to science 

 since the mid- 1800s, but their true signifi- 

 cance went largely unappreciated. Their 

 fossils have been recovered from Upper 

 Permian rocks (about 250 million years 

 old) in Russia, South America, China, and 

 Europe. Most specimens, however, have 

 come from South Africa, where farmers in 

 the dusty Karroo Basin often stumble 

 across bony remains weathering out of ex- 

 posed rocks. The local Afrikaans name for 

 them, handjietand dier (which is every bit 

 as annoying to spell as pareiasaur), refers 

 to their distinctive dentition and means 

 "animal with teethlike little hands." 



Pareiasaurs were among the largest ani- 

 mals of their time, but resemble nothing 

 alive today. Imagine a fat hippopotamus 

 with a thick tail. Shave off all its hair and 

 cover its back with little armor plates. 

 Now, stick some grotesque knobs all over 

 its skull. Finally, make it drag its belly 

 along the ground, with its legs sprawled 

 out sideways, like a lizard's or a turtle's. 

 The end result wouldn't look totally unlike 

 a pareiasaur. Aesthetically challenged to 

 say the least, these ponderous herbivores 

 have long been neglected by paleontolo- 

 gists, dismissed as an inconsequential evo- 

 lutionary dead end. (A colleague of mine, 

 Des Maxwell, branded them "history's 

 ugliest reptiles" and promptly switched to 

 working on dinosaurs instead.) Even their 

 name, pronounced "pariah-saur," seems to 

 invite such scom. 



As a less-discriminadng, first-year 

 graduate student, however, I persisted in 

 studying pareiasaur anatomy in more de- 

 tail. I discovered that although they resem- 

 bled bloated, oversized lizards in many re- 

 spects, pareiasaurs had already evolved 

 many of the characteristics of turtles. For 

 instance, all primitive reptiles completely 



lacked body armor. Early pareiasaurs, 

 however, had tiny bony plates embedded 

 in the skin over the backbone; in later 

 forms these plates spread out over the 

 sides and belly, and enlarged and fused 

 with one another to form a rigid cara- 

 pace — just like a turtle .shell. Also, most 

 eariy reptiles had long, slender bodies, 

 with twenty-five or more vertebrae in the 

 neck and back. Early pareiasaurs had short 

 bodies, with twenty vertebrae; later forms ^^ 

 were even shorter, with nineteen; and tur- HU 

 ties are stubbier still, with eighteen. Fi- -<^ 

 nally, moving from primitive reptiles to Q 

 early pareiasaurs to late pareiasaurs, the i— > 

 shoulder, pelvis, and limbs also became ^ 

 more and more turtlelike. The message r-h 

 was clear; turtles evolved from advanced '"' * 

 pareiasaurs. O 



If a pareiasaurian ancestry of turtles ^ 

 was so obvious, why hadn't anybody else 

 proposed it? Perhaps I had overlooked 

 some fatal weakness in the argument. As it 

 turned out, I wasn't the first person to 

 come up with the idea after all. William 

 Gregory, a curator at the American Mu- 

 seum of Natural History and one of this 

 century's paleontological greats, had pro- 

 posed the same idea almost half a century 

 ago, but his discussion of the supporting 

 evidence was vague. For instance, he 

 failed to emphasize that the similarities 

 shared by pareiasaurs and turtles were 

 found in no other primitive reptiles. His 

 views, therefore, were largely forgotten. 



What about more recent research? 

 People have long asserted that the turtle's 

 bizarre body plan is so highly modified 

 that all evidence of its ancestry has been 

 effectively obliterated. Therefore, many 

 recent workers assumed that only skull 

 features could reveal where turtles came 

 from. Thus, they overlooked aU the strik- 

 ing similarities between the bodies of 

 pareiasaurs and turtles. Scientists are no 

 more objective than other people; what we 

 see is heavily constrained by what we ex- 

 pect to see. My ignorance of established 

 dogma proved a godsend. Furthermore, I 

 was fortunate to have access to critical in- 

 formation unavailable to most previous 

 workers. The year before I began my stud- 

 ies, Eugene Gaffney, another curator at the 

 American Museum of Natural History and 

 an expert on fossil turtles, published a de- 

 tailed description of the 200-million-year- 

 old Proganochelys, the most primitive tur- 

 tle yet discovered. Knowing what this 

 turtle looked like was vitally important m 

 trying to figure out what its immediate an- 

 cestors looked like. Proganochelys re- 

 tained many features inherited from its 



