---
title: "A 13,800-Year-Old Alaska Camp Hints at Two Separate Routes Out of Asia"
description: "Diamond-shaped stone points at Shég’ Xdaltth’í’ match Siberian tools, not other Alaska sites, and the site may hold the oldest known shelter in the Americas"
author: "Luis Goa"
published: 2026-10-17T21:03:57Z
modified: 2026-10-18T03:55:08Z
url: https://rews.cc/a/a-13-800-year-old-alaska-camp-hints-at-two-separate-routes-o-d30abf
language: en
tags: ["archaeology", "alaska", "siberia", "ice-age", "anthropology", "science"]
publisher: "Rews (https://rews.cc)"
---

# A 13,800-Year-Old Alaska Camp Hints at Two Separate Routes Out of Asia

*Diamond-shaped stone points at Shég’ Xdaltth’í’ match Siberian tools, not other Alaska sites, and the site may hold the oldest known shelter in the Americas*

By Luis Goa · October 17, 2026 · https://rews.cc/a/a-13-800-year-old-alaska-camp-hints-at-two-separate-routes-o-d30abf

## In brief

- Alaska site Shég’ Xdaltth’í’ yielded over 80,000 archaeological specimens in its 13,800-year-old layer
- Diamond-shaped stone points there match tools from Siberia’s Nikita Lake and Berelekh, not other Alaska sites
- Researchers propose two separate Ice Age migrations carried different toolkits into Alaska
- Site may hold earliest known evidence of a built human shelter in North America, researchers say
- Findings published in Science Advances on October 14, 2026, led by Kelly Graf of the University of Kansas

One layer, one ridge above Alaska’s Tanana River, has produced more than 80,000 archaeological specimens dated to about 13,800 years ago. The site, Shég’ Xdaltth’í’ (“They were there” in an Alaska Native Athabaskan language), sits roughly 55 kilometers south-southeast of Fairbanks on land belonging to the Mendas Cha’ag and Nenana Tanana Athabaskan communities. A team led by Kelly Graf and Ted Goebel of the University of Kansas published the full description of the oldest layer on October 14, 2026, in [Science Advances](https://www.science.org/doi/10.1126/sciadv.aeh0802), as [Arkeonews reported](https://arkeonews.net/13800-year-old-alaska-camp-reveals-ancient-links-to-siberia/). Archaeologists first recorded the site in 2010; Graf’s team began digging there in 2013, continued annually until the pandemic, then returned in 2021, 2023 and 2025, work backed by $400,000 in US National Science Foundation grants, according to the [University of Kansas](https://news.ku.edu/news/article/bering-land-bridge-base-camp-reveals-what-life-was-like-nearly-14000-years-ago). The excavated footprint now covers at least 4,800 square meters.

## A point shape that doesn’t match its neighbors

The strongest evidence comes from stone, not bone. The oldest layer contains hafted bifaces, stone blades worked on both faces, that taper to a point at both ends and carry ground edges near the base, a finish that makes sense only if the stone was set into a handle or shaft. That diamond outline is not what turns up at other sites of Alaska’s early Nenana tradition, which favor triangular or teardrop points. It does turn up, in deposits of roughly the same age, at Nikita Lake and Berelekh in northeastern Siberia, and both of those Siberian sites, like Shég’ Xdaltth’í’, show no clear evidence of microblade technology, the small, standardized stone blade industry found at some other Beringian sites. “It’s the only site in Alaska that features this clear diamond shape,” Graf told the University of Kansas. “And the only other places at this time that have such points in a northern context are on the Russian side of the land bridge. It suggests people traveled across the Bering Land Bridge to and possibly from this location.”

A shared point style is weaker evidence than it looks. It shows that two groups of toolmakers, separated by the strait that now divides Alaska from Siberia, used the same design at roughly the same time. It does not show that the same individuals crossed, or that one group taught the other, or that both descended from a common population rather than converging on a workable shape independently. Graf and her co-authors go further than the tool comparison alone would support, proposing that Beringia’s Late Pleistocene record preserves two distinct population movements rather than one group doing different things in different places.

**Two hypothesized routes into Alaska** Proposed Late Pleistocene toolkit lineages, per Graf et al., Science Advances (2026)

| From | To | How |
| --- | --- | --- |
| Lena River basin, NE Asia | Microblade toolkit |  |
| Microblade toolkit | Swan Point, Alaska *(oldest unambiguous Americas site, ~14,000 years ago)* | carried east |
| Arctic NE Asia: Berelekh, Nikita Lake | Diamond hafted bifaces |  |
| Diamond hafted bifaces | Shég' Xdaltth'í', Alaska *(13,800 years ago)* | carried east |

![Two hypothesized routes into Alaska. Proposed Late Pleistocene toolkit lineages, per Graf et al., Science Advances (2026)](https://rews.cc/img/c61c325fe8a9c59f038912d205155d958edb7d4c.webp)

Based on [Graf et al., Science Advances](https://www.science.org/doi/10.1126/sciadv.aeh0802)

Under that reading, one stream carried microblade technology east from the Lena River basin, reaching sites such as Swan Point, dated to roughly 14,000 years ago and generally treated as the oldest unambiguously dated Ice Age site in the Americas, according to [Alaska state park records](https://dnr.alaska.gov/parks/oha/projects/swanpoint.htm). A second stream, originating further north in Arctic northeastern Asia and linked to Berelekh and Nikita Lake, carried the diamond-shaped bifaces that define Shég’ Xdaltth’í’ and what the team calls the Nenana Complex. For that hypothesis to hold, the two toolkits need to track two genetically distinguishable source populations, not one population’s regional variation. The stone record cannot settle that question; the researchers say ancient human DNA from the site’s deposits is the test still missing.

## A possible shelter, and what it cost to find it

Within the oldest layer, excavators mapped an oval patch of charcoal-stained sediment at least 3 meters across, containing a hearth about 70 centimeters in diameter, a shallow pit holding discarded bone, stone and charcoal, and what may be a posthole. No walls or roofing survived; the inference of a shelter, possibly a lean-to opening south, rests entirely on how that hearth, pit and posthole sit relative to each other. If the identification holds, it is the earliest evidence of a built human shelter yet found in North America, at 13,800 years. That claim rests on spatial arrangement, not surviving structure, and is the kind of inference later excavation at the same site, or at a comparably aged one, could overturn.

![A woodcut-style tundra landscape shows a sloping net shelter, a stone fire ring, and pits overlooking a winding river and distant mountains.](https://rews.cc/img/b63a95c3617b3a171f00744ed90c68f3b58357c8.webp)

A separate feature nearby, a dense concentration of charcoal, burned stone and animal bone, looks like an earth oven: heated rocks and soil used to slow-cook meat, a method that tenderizes tough cuts and helps render fat from bone. Researchers found broken bones consistent with marrow extraction and possible grease production, work that could have supported a fairly large family group and that the feature’s layout suggests was repeated across more than one visit.

## Needles, a pendant, and a long supply line

The oldest layer yielded 18 worked bone, antler or ivory objects: eight needle fragments, some retaining part of the eye through which thread was passed, an awl tip, several pointed tools, and a pendant about 2.9 centimeters long with part of a drilled suspension hole. The needles indicate hide-sewing capability sufficient for clothing and bags; the pendant is the camp’s only evidence of personal ornament. Researchers also recovered close to 6,900 bone specimens from the same layer, identifying bison, mammoth, horse, caribou, musk ox, arctic hare, wolves, foxes and several bird species including waterfowl and cranes, a range the team reads as evidence of more than one season of use. Presence in the bone pile does not by itself prove the animal was hunted; some of it may reflect other depositional processes the team is still working through.

Chemical sourcing of 25 obsidian artifacts traced the volcanic glass to the Batza Téna source, about 370 kilometers northwest of the site. The distance shows access to a resource well beyond the immediate area; it does not distinguish between the camp’s occupants traveling to get it themselves and getting it through exchange with other groups, and the researchers do not claim to know which. The oldest archaeological deposits here lie under more than 3 meters of sediment, which puts this layer at roughly [800 years older than the earliest Clovis sites](https://www.labrujulaverde.com/en/2026/10/this-alaskan-ice-age-camp-is-800-years-older-than-clovis-once-seen-as-americas-first-culture/) further south, the culture once credited as the first in the Americas.

![A woodcut-style scene shows a bone needle, cord, stone tools, and stitched hide overlooking winding rivers, forests, mountains, and a distant red mesa.](https://rews.cc/img/0b0ef78ef17c16a1a4ab0461fe2f63884a03bc67.webp)

One more anomaly: a wolf jawbone recovered from the site has a shape that falls outside the range of the ancient and modern wolf specimens the team used for comparison. That could point to another connection with northeastern Asian animal populations, and raises, without resolving, the question of whether the camp’s occupants kept early dogs. As with the human migration question, the researchers say the fix is genetic testing of the bone, not further argument from shape.

## A second paper, and an older idea revisited

Science Advances ran a companion commentary in the same issue, by Spencer Pelton of the University of Wyoming, who called Shég’ Xdaltth’í’ “remarkable because it is the oldest site in the Americas containing a large and rich assemblage of features, artifacts, and faunal refuse indicative of prolonged residential occupation, providing an unprecedented glimpse into the lives of North America’s first residents,” in [his commentary](https://www.science.org/doi/10.1126/sciadv.aem6534). Pelton’s piece revives an older hypothesis: that people moving into Beringia between roughly 20,000 and 15,000 years ago got stuck there, unable to move further south until retreating ice sheets opened a corridor. Both papers agree on the limits of what stone tools alone can show. A shared point shape, a shared absence of microblades, a shelter inferred from a hearth and a posthole: each is consistent with two separate migrations into Alaska, and each is also consistent with one population adapting its toolkit as it spread. The Tanana River sediment that preserved the camp has not yet preserved an answer to which. Ancient DNA, from the camp’s bones or the wolf jaw, is the next number to watch.

## See also

- [13,800-year-old Alaska camp links Ice Age hunters to Siberia](https://archaeologymag.com/2026/10/alaska-camp-links-ice-age-hunters-to-siberia/) — archaeologymag.com · Additional detail on the camp's stratigraphy and faunal assemblage
- [This Prehistoric Alaskan Camp Is Rewriting the History of Beringia](https://nautil.us/this-prehistoric-alaskan-camp-is-rewriting-the-history-of-beringia-1285804) — nautil.us · Narrative account of the excavation and the dwelling feature
- [13,800-Year-Old Alaskan 'Base Camp' Yields New Clues to Beringia Mystery](https://thedebrief.org/13800-year-old-alaskan-base-camp-yields-new-clues-to-ancient-beringia-land-bridge-mystery/) — thedebrief.org · Further coverage of the trans-Beringian tool comparison
- [Late Pleistocene site in Alaska yields evidence of a trans-Beringian cultural connection](https://www.eurekalert.org/news-releases/1147586) — eurekalert.org · AAAS press release accompanying the Science Advances paper