For 70 years, two large fossil bones sat in a University of Alaska museum drawer labeled as woolly mammoth — until radiocarbon dating, isotope analysis, and DNA testing revealed they came from two whales of two different species that somehow ended up 250 miles inland from the nearest coast, in a mystery the researchers say may never be fully solved

In 1951, gold miners at Dome Creek, a small settlement north of Fairbanks, donated a pile of fossil bones from their excavations to the explorer and naturalist Otto Geist, who in turn delivered them to the University of Alaska Museum of the North on 7 July of that year. Based on their general shape and where they had been found, two large bone discs from the pile were catalogued as the vertebral growth plates of woolly mammoths and given the specimen numbers UAMN3760 and UAMN3724. They sat in museum drawers for the next seventy years.

In 2022, the museum’s Adopt-a-Mammoth program, a crowdfunded project run by University of Alaska Fairbanks biogeochemist Matthew Wooller to radiocarbon-date as many of the museum’s mammoth fossils as possible, pulled the two bones for testing. The radiocarbon dates came back at 1,854 and 2,731 years old. If those dates had been right for mammoths, they would have been by far the youngest mammoth remains ever recorded on mainland Alaska, postdating the previously accepted extinction window by close to ten thousand years. Wooller’s team decided to confirm the result before announcing it.

What the bones actually were

The confirmation took two further rounds of testing. The first was stable isotope analysis, which compares ratios of carbon and nitrogen isotopes in the bone collagen to known ratios from different ecosystems. The values that came back from the two discs were inconsistent with any terrestrial herbivore, including a mammoth. They matched marine mammals. The second round was DNA sequencing. The DNA confirmed that the two bones were not from the same animal. One was from a North Pacific right whale, currently classified by the US National Oceanic and Atmospheric Administration as the rarest of the large whale species and endangered throughout its range. The other was from a minke whale, a much more common species. “The DNA evidence told us they were whales, but not even the same species of whale,” Wooller said in a University of Alaska Fairbanks news release. The findings were published in late 2025 in the Journal of Quaternary Science.

The bones themselves are epiphyseal plates, the disc-shaped growth surfaces that sit at either end of a vertebra in a young animal and fuse to the main bone with age. Their roughly circular shape and substantial size are why they had been mistaken for mammoth bones in the first place. Whale vertebrae and mammoth vertebrae look broadly similar at this stage of growth, particularly to a non-specialist examining them in the field.

One technical correction the team applied after identifying the bones as marine. The 1,854 and 2,731 year figures are uncorrected radiocarbon ages, which would be correct for a terrestrial animal but are systematically too old for a marine one. The carbon in ocean water is several hundred to over a thousand years older than the atmospheric carbon used as the standard reference for radiocarbon dating. Once this marine reservoir effect is taken into account, the true ages of the two whales drop to approximately 1,100 and 1,800 years.

The geographical problem

Identifying the bones as whales raised a different question. Dome Creek is roughly 400 kilometres, or 250 miles, from the nearest coastline. Both species, the North Pacific right whale and the minke whale, are open-ocean animals. Neither could have lived in the small inland rivers and creeks of central Alaska. The bones had to have arrived at Dome Creek by some other route. The Wooller team identifies three possible explanations in the paper, each with problems.

The first is an inland whale incursion through ancient inlets or large rivers. According to phys.org’s account of the research, the team considered whether the whales might have swum upriver via the Yukon and Tanana systems and died inland. This is judged unlikely. The North Pacific right whale is a plankton feeder that requires the kind of open-ocean food sources not available in interior rivers, and even at the river volumes of two thousand years ago, the waterways around Fairbanks were probably too small to accommodate a whale of that size. Wayward cetacean strandings far from typical habitat do occur, but a single live whale ending up 400 kilometres up a freshwater river system is at the extreme end of the literature, and accounting for two whales of two different species this way is harder still.

The second explanation is human transport. Whale bone was a valued material in coastal Alaskan cultures, used for tools, plates, structural elements, and ritual objects. Archaeological records document extensive trade and transport of whale parts inland from coastal communities. Patrick Druckenmiller, a UAF palaeontologist and co-author on the study, told UAF’s Geophysical Institute that the bones in question show no obvious signs of being modified for use as tools or platters. That does not rule out human transport, since unmodified bones could have been carried inland for other purposes or simply not yet shaped, but it also does not provide direct evidence for the theory.

The third explanation, which the authors describe as less dramatic but at least as plausible, is a museum accessioning error. Geist collected specimens from across Alaska throughout the 1920s to the 1960s, including from Norton Bay on the western coast, where whale remains are common and easy to come by. According to the Geophysical Institute account, on the same day the Dome Creek bones were being processed, 7 July 1951, the museum was also cataloguing more than twenty fossils Geist had collected from Norton Bay. The whale vertebrae may have been put in the wrong bin. “It is possible they got juggled,” Wooller said. The white paint markings on the bones themselves, showing the year and sample number, were applied by Geist or museum staff and have stayed legible across the decades, but the underlying provenance is no longer independently verifiable.

Why the question may stay open

The three hypotheses are not mutually exclusive, and the researchers do not endorse any of them as a settled answer. The team’s framing, in the published paper, is that the bones are now known to be whale, the original radiocarbon dates are real, the location they were catalogued from is well documented, and the route from coastal whale to interior museum drawer is no longer recoverable in detail. “Ultimately, this may never be completely resolved,” Wooller and his co-authors wrote.

The deeper context, which the Wooller group also flag, is that historical museum collections are full of specimens that were identified using the methods available at the time of collection and have not been re-examined since. The radiocarbon dating and stable isotope and DNA tools used to redescribe UAMN3760 and UAMN3724 are recent enough that the original catalogers in 1951 had no access to them. As ScienceAlert noted in its coverage of the study, the broader value of the Adopt-a-Mammoth project, beyond any specific finding, is that it produces a steadier flow of redescribed and properly dated specimens from collections that have been quietly sitting in storage for decades.

The two specimens have been re-catalogued. The mammoth timeline for mainland Alaska is unchanged. The two whales have been added to the museum’s records, with full species identification and dating, and are now available for use in future research on Late Holocene marine mammals in the North Pacific. How they got to Dome Creek is the part that may stay unsettled.