New Ways to The New World

By ANDREA DORFMAN

It is the Americas' epic immigration saga, long taught in schools and enshrined in popular books. At the end of the last Ice Age about 12,000 years ago, brave Siberians walked across the Bering Sea land bridge, then edged their way south via a newly opened corridor in the ice and fanned out in all directions. Within 500 years, their descendants had settled most of the hemisphere, from the Arctic Circle to the tip of South America. Alas, as archaeologists have learned by digging up and down the Americas, this engaging tale may be wrong.

The latest evidence against the old story was unveiled last week in Philadelphia during the annual meeting of the Society for American Archaeology. Joseph McAvoy of the Nottoway River Survey and his colleagues disclosed that an ancient campsite known as Cactus Hill, 45 miles south of Richmond, Va., has been conclusively dated at around 18,000 years old. That predates the accepted timing for the opening of that crucial ice-free corridor and bolsters the theory that the earliest Americans came by sea, possibly even from across the Atlantic rather than from Asia. "If the dates hold up, and I think they will," says archaeologist Dennis Stanford of the National Museum of Natural History in Washington, "this is probably some of the oldest material in North America, if not the entire New World."

For decades, 11,200-year-old stone spear points from a site in Clovis, N.M., have been held to be the earliest evidence of settlement in the hemisphere--and many archaeologists have been loath to give up this "Clovis first" model. But since the 1970s, it has been challenged by the discovery of still older sites on both sides of the continent, most notably a 17,000-year-old rock shelter in Meadowcroft, Pa.

Now Cactus Hill presents still more corroboration. Taking its name from the prickly pears that grow at the site, it was discovered in 1988 by a sharp-eyed farmer named Harold Conover, who alerted researchers to some curious stone tools he had spotted in road sand dug up from an old pit nearby. In 1989, McAvoy's team began excavations, now sponsored by the National Geographic Society and the state of Virginia. So far, the team has unearthed a variety of Paleo-Indian stone tools shaped for hunting, butchering and processing game; charred bones of mud turtles, white-tailed deer and other mammals; and bits of charcoal left over from hunting parties' cooking this prey.

Radiocarbon dating and other techniques indicate the campsite was occupied as long as 5,000 years before the Clovis culture appeared. Calling the results "unequivocal," McAvoy says they should "terminate the debate over whether Clovis was first or not." The Meadowcroft rock shelter's chief investigator, archaeologist James Adovasio of Mercyhurst College in Erie, Pa., agrees. "This is another indication that people were running around North America earlier than 13,000 years ago," he says.

If so, how could they have got here? One growing possibility, long thought heretical: by boat along the eastern and western coasts of the Americas. A 12,500-year-old settlement in Monte Verde, Chile, for example, seems to have been reached most easily by water. The lack of any evidence of shipbuilding doesn't dissuade Adovasio. Says he: "You had southeast Asians sailing to Australia more than 50,000 years ago."

The most startling idea is raised by Stanford, who says the Cactus Hill tools resemble even older ones found in Spain and France. He and archaeologist Bruce Bradley of Cortez, Colo., propose that the first people to reach the Americas worked their way across the Atlantic from the Iberian Peninsula some 17,000 to 18,000 years ago.

For now, few scientists are willing to go so far. "I think people did have the capacity to sail across the Atlantic," says Adovasio, "but I still think 99.9% of the peopling of the Americas occurred through the interior or along the coast from the Bering Sea." Still, he leaves a tantalizing 0.1% to begin some new mythmaking.