North and South America were first populated by three waves of migrants from Siberia rather than just a single migration, say researchers who have studied the whole genomes of Native Americans in South America and Canada.
Some scientists assert that the Americas were peopled in one large migration from Siberia that happened about 15,000 years ago, but the new genetic research shows that this central episode was followed by at least two smaller migrations from Siberia, one by people who became the ancestors of today's Eskimos and Aleutians and another by people speaking Na-Dene, whose descendants are confined to North America. The research was published online Wednesday in the journal Nature.
The finding vindicates a proposal first made on linguistic grounds by Joseph Greenberg, the great classifier of the world's languages. He asserted in 1987 that most languages spoken in North and South America were derived from the single mother tongue of the first settlers from Siberia, which he called Amerind. Two later waves, he surmised, brought speakers of Eskimo-Aleut and of Na-Dene, the language family spoken by the Apache and Navajo.
But many linguists who specialize in American languages derided Greenberg's proposal, saying they saw no evidence for any single ancestral language such as Amerind. "American linguists made up their minds 25 year ago that they wouldn't support Greenberg, and they haven't changed their mind one whit," said Merritt Ruhlen, a colleague of Greenberg, who died in 2001.
The new DNA study is based on gene chips that sample the entire genome and presents a fuller picture than earlier studies, which were based on small regions of the genome such as the Y chromosome or mitochondrial DNA. Several of the mitochondrial DNA studies had pointed to a single migration.
A team headed by David Reich of the Harvard Medical School and Andres Ruiz-Linares of University College London report that there was a main migration that populated the entire Americas. They cannot date the migration from their genomic data, but they accept the estimate by others that the migration occurred around 15,000 years ago.
This was in the window of time that occurred after the melting of great glaciers that blocked passage from Siberia to Alaska, and before the rising waters at the end of the last ice age submerged Beringia, the land bridge between them.
They also find evidence for two further waves of migration, one among Na-Dene speakers and the other among Eskimo-Aleut, again as Greenberg predicted. But whereas Greenberg's proposal suggested that three discrete groups of people were packed into the Americas, the new genome study finds that the second and third waves mixed in with the first. Eskimos inherit about half of their DNA from the people of the first migration and half from a second migration. The Chipewyans of Canada, who speak a Na-Dene language, have 90 percent of their genes from the first migration and some 10 percent from a third.
It is not clear why the Chipewyans and others speak a Na-Dene language if most of their DNA is from Amerind speakers. Dr. Ruiz-Linares, who is also a medical doctor, said a minority language can often dominate others in the case of conquest; an example of this is the ubiquity of Spanish in Latin America.
If the genetics of the early migrations to the Americas can be defined well enough, it should in principle be possible to match them with their source populations in Asia.
Greenberg had argued on linguistic grounds that the Na-Dene language family was derived from Ket, spoken by the Ket people in Siberia's Yenisei valley. But Mr. Reich said there was not yet enough genomic data from Asia or the Americas to make these links. His samples of Na-Dene and Ket DNA did not match, but the few Ket samples he had may have become mixed with DNA from people of other ethnicities, so the test, in his view, was inconclusive.
The team's samples of Native American genomes were drawn mostly from South America, with a handful from Canada. Samples from U.S. tribes could not be used because the existing ones had been collected for medical reasons, and the donors had not given consent for population genetics studies, Dr. Ruiz-Linares said.
Native Americans in the United States have been reluctant to participate in inquiries into their origins. The Genographic Project of the National Geographic Society wrote recently to all federally recognized U.S. tribes asking for samples, but only two agreed to give them, said project director Spencer Wells.
Interracial marriage -- or admixture, as geneticists call it -- may have distorted earlier efforts to trace ancestry because subjects assumed to be American may have had European or other DNA admixed in their genomes. Mr. Reich and his colleagues have developed a method to define the racial origin of each segment of DNA and have found that, on average, 8.5 percent of Native American DNA belongs to other races.