First Americans arrived 2500 years before we thought

It's time to rewrite the story of how Stone Age explorers from Asia crossed over into the Americas and colonised the continents.

The Clovis people were leading candidates for the title of first Americans. But a hoard of tools newly uncovered in Texas suggests the land was inhabited several thousand years before the reign of the Clovis culture.

When the people who built the Texan tools migrated, ice sheets would have made travel by land difficult. This lends strength to the hypothesis that the Americas were colonised by sea, not land.

Who the first Americans were, where they came from and when they arrived are contested issues among archaeologists. One favoured theory, known as "Clovis first", says that during the last Ice Age, people from Asia followed herd animals across a land bridge connecting Siberia to Alaska and established the first settlements in North America. The Clovis culture is characterised by pointed stone tools.

But recent finds of artefacts that pre-date the Clovis, including this new one in Texas, have challenged the Clovis-first hypothesis.

Flakes and stones

The new hoard contains 15,528 items, the largest group of pre-Clovis stone objects ever found. It includes 56 well-preserved tools amongst many stone chips, flakes and fragments that probably broke off other tools.

"What we have found is evidence of early human occupation dating back to 15,500 years ago, 2500 years older than Clovis," says Michael Waters of Texas A&M University in College Station, lead author of the study.

Waters and his team discovered the primitive toolkit in a well-preserved layer of soil at Buttermilk Creek in central Texas. Directly above it lay another, distinct layer dating from the Clovis era.

The objects are clearly shaped by human hands, but less sophisticated than Clovis tools - the team describe them as prototypes. The hallmark of Clovis technology is a carefully chiselled, oval-shaped stone with thin razor-sharp edges and a notch in the bottom for hafting to a spear or knife handle. In contrast, the newly discovered tools are not well-shaped, lack notches and are lighter than Clovis tools. Waters thinks that descendants of their makers may have later invented Clovis technology.

Difficult dating

Establishing that the objects were indeed older than the Clovis proved a challenge. The team did not find enough organic matter at the site to pinpoint the age of the tools with radiocarbon dating. Instead they relied on a newer and slightly less accurate technique called optically stimulated luminescence, which uses light to free electrons trapped in minerals.

Others agree the find is significant. "This looks to me like a really solid example of archaeology that is older than dates people associate with Clovis," says Douglas Bamforth of the University of Colorado at Boulder. "They have done a great job of documenting the age of the sediment."

Dennis Jenkins of the University of Oregon in Eugene is similarly impressed, but points out that the tools could have shifted through the ages. "Nobody will argue these artefacts aren't real, but the question is whether they were really found exactly where they belonged or whether they settled from above."

Quiet grave

Burrowing rodents, plant roots and geologic activity all create cracks and voids in soil. The artefacts could have slipped over time through such gaps from the higher Clovis layer to the older Buttermilk Creek layer. But Waters and his team argue this is not likely to have happened here.

Firstly, the site is not especially geologically active and the team did not find any cracks large enough for objects to sift through. Secondly, if the earth had shifted, allowing the artefacts to move about, the changes would show in the magnetic signatures of different layers of soil; the team analysed the magnetic record, but found no such signs of disturbance.

Finally, the team showed that they could piece stone flakes together like pieces of a 3D jigsaw puzzle, and the pieces that fitted together always came from a single layer of earth. In other words, the fragments had not moved from their original burial site.

Paradigm shift

The new finds also suggest that the bridge between Asia and America was not the only route into the Americas. Fifteen thousand years ago, people in Siberia could not easily have crossed to Alaska and down into North America because the major ice sheets at the time were fused, prohibiting travel through North America after crossing the bridge. Instead, whoever made the stone tools at the Buttermilk Creek site may have journeyed to the New World by sea.

"I think we are on the edge of a paradigm shift now," says Waters. "We're past the Clovis-first model. We have robust evidence of people here before Clovis that is in a secure geological context and well-dated. Now we can seriously sit down and develop a new model for the peopling of the Americas."

Bamforth echoed his opinions. "It might be time to stop talking about Clovis-first altogether."

Journal reference: Science, DOI: 10.1126/science.1201855