The question of when and how humans made it to the Americas is one of the most heated debates in paleontology – a field particularly known for scientific conflict. The discovery of mammoth bones that appear to have been made into tools could push the arrival date earlier than almost anyone has imagined. However, after the debunking of a paper attributing marks on even older bones to butchery, the claim must overcome initial skepticism if it is to be accepted.

Crossing the Pacific Ocean represents one of humanity’s greatest journeys – indeed some would say one of our greatest achievements of any sort, given the technology available. Indeed, every explanation for how it could have been done has problems. Narrowing down the timing might at least rule out some options.

This requires the finding of the oldest evidence of a human presence in the Americas, or at least something close to it. In a new paper in Frontiers in Ecology and Evolution, scientists at the University of Texas at Austin describe mammoth bones found in New Mexico’s Rio Puerco canyon which they consider the most convincing evidence of early arrival yet. 

However, while most discoveries considered candidates for the earliest signs of humanity in the Americas are thought to be 21,000 to 24,000 years old, with a few out to 30,000, this one is around 37,000 years old.

“What we’ve got is amazing,” said Professor Timothy Rowe in a statement.  “It’s not a charismatic site with a beautiful skeleton laid out on its side. It’s all busted up. But that’s what the story is.”

Rowe and co-authors believe what the Rio Puerco site lacks in organization, it makes up for in its extraordinary combination of different pointers to human presence. These include what appear to be bone knives with edges worn but initially suited to cutting, signs of controlled fire, and bones that suffered blunt force trauma. 

In three-quarters of the sites, many bone flakes were struck either parallel or perpendicular to the grain, consistent with tool making and unlike bone collections broken by scavengers or geological forces.

None of the bones show toothmarks from carnivore scavenging, possibly evidence humans kept other animals away.

Collagen extracted from mammoth bones at the site has previously been dated to between 36,250 and 38,900 years ago, so even the younger end of that age range is far earlier than any other accepted sign of human presence in the New World.

The oldest confirmed tools found in the Americas are from the so-called Clovis culture, and date to around 16,000 years ago. Where once the Clovis people were thought to be the first arrivals in the Americas, using a land bridge and passage between ice sheets,  increasing evidence has emerged of pre-Clovis cultures. However, it is one thing to accept people reached the Americas a few thousand years before Clovis, quite another to push that back by 200 centuries, as this would do.

Some genetic evidence suggests the Americas were settled by two populations, one of East Asian ancestry, followed much later by a group that provided the majority of the pre-European genome.

Rowe’s expertise lies in dinosaurs. However, when a neighbor pointed out a mammoth tusk sticking out a hillslope on Rowe’s property in 2013 the opportunity to diversify was too good to miss. Digging revealed a smashed mammoth skull and other bones damaged in ways Rowe and colleagues consider more likely to be deliberate than accidental. CT scans and environmental scanning electron microscopy revealed fracture networks on bones resembling those on cow bones cut as tools rather than broken by accident. 

Sediments at the site show evidence of long-lasting fires, consistent with human activity for cooking or warmth, rather than lightning-struck wildfires. The presence of fish well above the nearest river, and an array of other small animal carcasses further support the theory humans cooked here, rather than the site representing a random location for animal deaths.