Retelling the tale of the two-legged snake.

In 1978, the late geologist George Haas, working near Jerusalem, found the fossil of a sinuous, meter-long creature with two stubby rear legs. He called it a lizard. When Michael W. Caldwell and Michael S.Y. Lee reexamined the bones recently, they came to a different conclusion.

"It's the missing link between the snake and the lizard," says Lee, a paleontologist at the University of Sydney in Australia. The reclassification challenges the dominant theory that snakes evolved solely from burrowing lizards, since the Israeli fossil represents an aquatic creature. The two scientists posit in the April 17 Nature that modern snakes descended from giant sea lizards called mosasaurs, which became extinct with the dinosaurs.

Caldwell and Lee did a bone-by-bone comparison between the 97-million-year-old fossil, Pachyrhachis problematicus, and other animals. The fossil resembles a snake in that it has mobile jaws and a narrow skull that fully encloses the brain. The fossil's an kles have two distinct bones that in land lizards are fused.

Both the fossil and the mosasaurs have a pelvic girdle and tiny rear limbs. Some living snakes, such as the boa constrictor, contain traces of a pelvis and bony knobs that could be the vestiges of limbs. Even though the fossil's centimeters-long rear legs are fully formed, they were probably too small to serve any purpose, suggesting that over generations the creatures were slowly losing their legs.

Caldwell, a paleontologist at the Field Museum in Chicago, says it's conceivable that two lines of snakes developed, one from burrowing lizards and another from aquatic lizards. For now, he considers both evolutionary theories equally plausible, but he su spects that as scientists find more fossils, the weight of the evidence will tip toward a marine origin. Only two P. problematicus fossils have been found, both by Haas and both in the same location. The other one has a crushed skull and pelvis.

Nicholas C. Fraser of the Virginia Museum of Natural History in Martinsville agrees that the reexamined fossil more closely resembles a snake than a lizard. He also says that Lee and Caldwell's argument for a marine origin is persuasive but not conclusive . The link between the early snake and the mosasaurs may be coincidental, he warns. For example, species like bats and pterosaurs share many features, but they evolved independently.