First Americans Arrived Around 18,000 Years Ago, Genetic Evidence Suggests
New World Ancestors Lose 12,000 Years
July 25, 2003
Scientists studying the genetic signatures of Siberians and American Indians have found evidence that the first human migrations to the New World from Siberia probably occurred no earlier than 18,000 years ago.
The new estimate undermines arguments for colonization as far back as 30,000 years ago, but reinforces archaeological findings and a linguistic theory that most American languages belong to a single family called Amerind.
The genetic evidence fits neatly, for example, with the discovery of a human campsite in Chile, which is apparently 15,000 years old, and with the well-established presence of big-game hunters in North America, starting 13,600 years ago. The few sites with possibly older human traces have yet to gain wide acceptance among scientists.
By studying the DNA of living Siberian and American Indian populations, geneticists had previously been able to see traces of at least two early migrations from Siberia. But it has been hard to put a date on when the first people set foot in the Americas, for lack of a suitable marker in the Y chromosome.
After much search, a team of geneticists has now detected a change in the DNA sequence of Siberian men's Y chromosomes that took place just before the first of the two migrations into the Americas. They estimate that the DNA change, called M242, occurred 15,000 to 18,000 years ago, meaning the Americas must first have been occupied after that date. The DNA change is not in a gene and makes no known difference to the men who carry it.
The new result, to be published in the American Journal of Human Genetics, is by Dr. Mark Seielstad of the Harvard School of Public Health, Dr. R. Spencer Wells of the University of Oxford and other colleagues.
The migration was probably by land because at that time the world's sea level was much lower and a land bridge, known as Beringia, stretched across what is now the Bering Strait between Siberia and Alaska. Also, people bearing the same genetic marker, called M3, live on either side of the former bridge, suggesting it was the means of passage.
Beringia sank beneath the waves some 11,000 years ago as the glaciers of the last ice age melted. The second migration seen by the geneticists seems to have occurred some 8,000 years ago and was presumably by boat, as the land bridge had long since vanished.
The date based on the new marker is important because it sets an earliest limit on the colonization of America, something that archaeologists find hard to do because they cannot be sure there are not sites they may have missed.
Hitherto some archaeologists have argued that people reached the Americas as long as 30,000 years ago. This date received some genetical support last year in a study by Dr. Douglas Wallace, now of the University of California at Irvine, who matched up male migrations from Siberia with the female migrations that he and colleagues had worked out earlier. The female migrations are traced by analyzing a genetic element in every cell called mitochondrial DNA.
Based on the mitochondrial DNA of the women descended from those in the first migration, Dr. Wallace estimated it occurred 20,000 to 30,000 years ago. Dr. Spencer said in an e-mail message that mitochondrial DNA was hard to date accurately and often gave dates that were too old. The Y chromosome is a better genetic clock, if a suitable marker can be found, he said.
Dr. Wallace did not respond to e-mail requests for comments.
The new date derived by Dr. Seielstad and Dr. Spencer may strengthen the hand of linguists who argue that all American languages fall into three groups, known as Amerind, Na-Dene and Eskimo-Aleut, with Amerind being by far the largest. Most linguists dispute that classification, saying languages change too fast to allow any very ancient relationships to be discerned. But if the first humans arrived in the Americas only 18,000 years ago, efforts to find links between present languages may seem more plausible.
"If they entered more recently, it is not such a stretch to say you can see a linguistic relationship," Dr. Wells said.
The new archaeological results seem compatible with the younger date adduced by the geneticists. Radiocarbon dating revealed that a occupation site in Siberia was only 13,000 years old and thus too recent to be a critical link in the first migrations, as had been supposed.
The site on the Kamchatka Peninsula of Russia, previously dated at 16,800 years old, was thought to be a way station at the western edge of Beringia, a point of departure for migrants either across the frozen land or by sea along the coasts. The new research challenges the conventional idea that this was the specific site from which people crossed into America, but does not exclude the possibility that they did so from other sites.
Researchers, led by Dr. Ted Goebel of the University of Nevada at Reno, reported the redating of the Siberian site at Ushki Lake on the Kamchatka Peninsula in today's issue of the journal Science. The other authors were Dr. Michael R. Waters of Texas A&M University and Dr. Margarita Dikova, an archaeologist and widow of Dr. Nikolai Dikov, who discovered the site in 1964.
The initial radiocarbon analysis was apparently based on contaminated samples, the researchers said. The 13,000-year-old date, nearly 4,000 years younger than previously thought, effectively removed the site as a way station for the first migrants to America, they concluded.
For most of the last century, the peopling of America was a story of big-game hunters trekking across the Bering land bridge in the last ice age, spreading across North America and within 1,000 years or so reaching the tip of South America. Those who left the most durable traces, fluted projectile points, were the Clovis people, named for the town in New Mexico where their artifacts were first uncovered.
The journal quoted Dr. David J. Meltzer, an archaeologist at Southern Methodist University, as saying the new finding "removes what was, until now, the critical link in the chain connecting Clovis to Siberia."
When people first occupied the Ushki Lake site, Clovis hunters had already been killing mammoths in North America for some 600 years and groups of hunters had left their mark at Monte Verde, Chile, 3,000 years earlier. Radiocarbon dates are lower than calendar dates and they become increasingly so the farther back one goes in time.
If the Ushki site is only 13,000 years old, Dr. Goebel said, the oldest place in the Bering region with human traces now is Broken Mammoth, a 14,000-year-old site in central Alaska.
"It means we have even less evidence than we had before," Dr. Goebel said.