Science

A 13,800-year-old Alaska camp holds the earliest sign of a house in North America

Nadia Okonkwo
Add us on Google

On a windy bluff above the Tanana River in interior Alaska, archaeologists have found the floor of a shelter that people built nearly fourteen thousand years ago. A ring of charcoal-stained earth surrounds a hearth, a small refuse pit and the stain of a wooden post, the traces of what was most likely a lean-to open to the south.

If that reading holds, it is the earliest known case of a habitation structure in North America. It also changes the picture of the first Alaskans: not only small hunting parties passing through, but families who settled in, cooked for a crowd and came back to the same spot for generations.

The site is called Shég’ Xdaltth’í’, which means “They were there” in the Alaska Native Athabascan language. It lies about 55 kilometers south-southeast of Fairbanks, on the traditional lands of the Mendas Cha’ag and Nenana communities. A team led by Kelly Graf and Ted Goebel of the University of Kansas excavated it with the Healy Lake Tribal Council and the Nenana Native Association.

“I expected we’d find hunting tools and knapping debris surrounding a hearth or two, but I had no idea that we’d stumble into a base camp with tens of thousands of artifacts,” Graf told Live Science. “It was a total surprise!”

The oldest layer alone holds more than 80,000 archaeological specimens, packed into an excavation block of just 25 square meters. Among them are eight fragments of eyed bone needles, an antler pendant with a drilled hole, a beveled ivory point and nearly 7,000 animal bones.

A shelter, an oven and a sewing kit

The shelter is an oval patch of greasy, charcoal-darkened sediment at least three meters across. Inside it sat a hearth about 70 centimeters wide and a shallow pit about 80 centimeters across, filled with bone, stone flakes and charcoal. A possible post mold, roughly 10 by 18 centimeters, marks where a pole once stood.

A few meters away, the team found a pile of discarded bones and something more ambitious: an earth oven. It is a burned pit more than 1.2 by 1.7 meters, packed with charcoal and 230 fire-cracked rocks, where cooks buried meat with heated stones to slow-cook it. The authors suggest it fed a large family group, and its radiocarbon dates span roughly two centuries, so people likely fired it up again and again.

The menu was broad: bison, horse, mammoth, caribou and musk ox, but also arctic hare, foxes, a lynx and at least ten kinds of birds. Waterfowl point to summer hunting, and fur-bearing animals to winter trapping. Together, the authors argue, they describe a base camp used across seasons by a group larger than a hunting party.

“Early Alaskans knew their surroundings well,” the researchers wrote.

Blades from Siberia and a wolf that matches no known population

The stone tools tie the camp to the far side of the Bering Strait. Six finished blades have a distinctive bipointed, diamond-shaped outline with ground-down stems, a sign they were fixed to handles or shafts. That shape is unknown elsewhere in central Alaska, where blades of this age are triangular or teardrop-shaped. It does turn up at Nikita Lake and Berelekh, two sites of the same age in the Yana-Indigirka Lowland of northern Siberia.

The raw materials show how far these people ranged: every piece of obsidian came from Batza Téna, a volcanic source 370 kilometers to the west-northwest.

Then there are the canids. The camp holds at least 15 bones from wolf-like animals. A shape analysis of one jaw does not match dire wolves, Ice Age Beringian wolves or modern wolves from the northern United States, which suggests a distinct population. Genetic studies indicate that early North American dogs descend from Siberian wolves drawn to human camps, so the authors raise the possibility that the people here kept dogs. They also stress that early dogs and wolves can be impossible to tell apart from bones.

How they did it

The team mapped nearly 51,000 objects in three dimensions as they dug. That record lets them read features such as the shelter outline: the spread of burned rock, bone and flakes around the hearth traces the edges of a living space that left no walls behind.

Dating rests on 66 radiocarbon dates, 50 of them from the oldest layer, mostly on willow charcoal and animal bones. A Bayesian statistical model, which combines the dates with the order of the layers, places the main occupation between 13,850 and 13,500 years ago, with hints of visits as early as about 14,050 years ago. Two shorter occupations followed, at about 13,000 and 12,700 years ago.

To identify crushed fragments that no longer look like any particular bone, the researchers used ZooMS, a technique that reads collagen proteins to tell species apart. It identified 178 of the 373 bones assigned to an animal group so far.

Why the story of the first Americans runs through this valley

The Bering Land Bridge, the plain that joined Siberia and Alaska during the last ice age, is where the ancestors of Indigenous Americans are thought to have crossed. Recent studies suggest people had entered the Americas by around 21,000 years ago, but the traces they left in Beringia are scarce and usually small: a hearth, some flakes, a few bones from a short hunting stop.

Shég’ Xdaltth’í’ sits in the densest cluster of early sites on the continent. Nineteen of the 28 known Beringian occupations from before or during the Clovis era, named after the culture that spread across North America about 13,000 years ago, are packed into a patch of east-central Alaska and southwest Yukon slightly larger than the US state of Maryland. The oldest of them, Swan Point in the same valley, is about 14,200 years old and used a different toolkit built around tiny stone microblades.

That contrast is the paper’s boldest idea. The authors propose that the diamond-shaped blades of Shég’ Xdaltth’í’, Berelekh and Nikita Lake mark one migration out of arctic northeast Asia, while the microblade makers of Swan Point came from the subarctic Lena River basin. Two toolkits, in other words, could mean two peoples.

The camp also shares a list of traits with later Paleoindian cultures south of the ice sheets, from eyed needles and ivory points to red ochre, personal ornaments and mammoth hunting. “The co-occurrence of these cultural features at the single site of Shég’ Xdaltth’í’ is truly notable,” the researchers wrote.

What it doesn’t settle

The team itself hedges the house claim. The authors call the shelter “the earliest potential case of a habitation structure in North America” and describe it as ephemeral. There are no storage pits or sunken house floors like those at Ushki in Kamchatka or Mal’ta in Siberia, and the main feature still runs into the excavation wall, so its full outline is unknown.

Nor is this the oldest site in Beringia, as the authors state in their abstract. Swan Point is older, and a few places, such as the Yana site in Siberia and Bluefish Caves in Yukon, hint at people in the region far earlier.

The dog question remains open, and so does the two-migrations hypothesis. Linking the camp’s toolmakers to the populations geneticists have reconstructed will require ancient genomes, and no human remains have been found at the site. The statistical model also compresses the main layer into about 280 years, which may blur several separate visits into one.

Movement across the region is harder to prove. “It would be hard to pin down whether or not they were moving back and forth during a lifetime, but at least we can say there is a connection,” Graf told Live Science.

Common questions about the Alaska site

What is the oldest house in North America?

The strongest candidate is now a shelter at Shég’ Xdaltth’í’ in interior Alaska, dated to about 13,800 years ago. Its excavators call it the earliest potential habitation structure in North America: a lean-to around a hearth, not a solid house.

When did people first reach Alaska?

Swan Point, in the same Tanana Valley, shows people present by about 14,200 years ago. A handful of sites hint at earlier visits, and people had reached the Americas by around 21,000 years ago, which implies an earlier passage through Beringia.

Did the first Americans have dogs?

Possibly. Genetic evidence suggests North America’s earliest dogs came from Siberian wolves, and the wolf-like animals at Shég’ Xdaltth’í’ belonged to an unusual population. Early dogs and wolves are hard to tell apart from bones, so the question remains open.

The study was published on 14 October 2026 in Science Advances, after excavations that began at the site in 2016. Its authors include researchers from France’s CNRS, Germany’s Max Planck Institute of Geoanthropology and the University of Alaska Museum of the North.

The next answers will come from the bones. The team plans to compare the canids with ancient Siberian wolves, run genetic tests, finish protein identification of the fragments and study seeds and plant remains from the hearths to pin down the seasons of each stay. The eastern edge of the shelter still lies unexcavated, inside the trench wall.

Reference: Graf et al., “Highlights from an early base camp in Beringia 13,800 years ago: Shég’ Xdaltth’í’ and the Late Pleistocene peopling of Alaska,” Science Advances, 2026. DOI: 10.1126/sciadv.aeh0802

Tags: , , , , ,

Add us on Google

Discussion

There are 0 comments.