A limestone cave near Kjøpsvik in northern Norway preserved something glaciers usually destroy: a layered record of an entire coastal Arctic community from roughly 75,000 years ago. More than 6,000 bone fragments revealed polar bear, walrus, bowhead whale, reindeer, seabirds, freshwater fish and a collared lemming never previously recorded in Scandinavia.
The research was published in 2025, but the first bone was discovered in 1991 and the major modern excavations took place later. The PNAS paper states that the collared lemming is extinct in Fennoscandia and had not previously been found in Scandinavia. The species itself still lives farther east.
Those distinctions do not diminish the discovery. They make clear what the cave actually provides: the oldest known diverse animal community from the European Arctic during this warmer interval of the last glacial period, assembled from bones, sediments and ancient DNA.
The research began with a bone found in 1991
Arne Qvamgrotta is a branch of the Storsteinhola cave system at 68.1 degrees north, just inside the Arctic Circle. Limestone mining and tunnel work exposed the deposit. On 8 August 1991, geologist Stein-Erik Lauritzen found a polar bear rib that preliminary work dated to at least 70,000 years old. The University of Bergen’s history of the project describes how that discovery started more than three decades of work.
A University of Oslo-led team returned for larger excavations in the early 2020s. The researchers’ public account identifies the 2021 and 2022 field seasons, while the study appeared in PNAS on 4 August 2025. The 2025 date therefore refers to publication, not the discovery or excavation of the remains.
Seventy-five thousand years is a useful midpoint
The bones do not all carry a neat 75,000-year timestamp. The team combined radiocarbon limits, optically stimulated luminescence, uranium-thorium dating, cosmogenic burial dating and estimates from evolutionary family trees. Together, the evidence places the main assemblage in Marine Isotope Stage 5a, most likely between about 85,000 and 71,000 years ago.
This was the Odderade interstadial, a relatively mild phase within the Weichselian glacial period. Ice had retreated enough to leave stretches of the northern Norwegian coast exposed. Because later glaciers and meltwater normally scrape away or rearrange old deposits, intact organic remains from before the Last Glacial Maximum are exceptionally scarce in Fennoscandia.
The cave’s value comes from the sedimentary context as much as from any single bone. The fragments were concentrated in a stratified unit rather than scattered without order. Multiple dating approaches, the fauna and the condition of the bones led the authors to treat it as one ecological window, although they acknowledge that material could have accumulated over thousands of years.
That uncertainty matters because “community” here does not mean every animal stood on the coast at the same moment. It means their remains came from a coherent interval and record a compatible environment, not that the cave captured a single catastrophic event.
DNA recovered an ecosystem from mostly anonymous fragments
The assemblage was highly broken up. Traditional comparative osteology, which identifies an animal from the shape of a tooth or bone, could assign only 350 fragments, about 5.7 per cent of the collection, to a family or a more precise group. The researchers therefore added bulk-bone metabarcoding, extracting short DNA sequences from batches of fragments and matching them against genetic reference libraries.
Together, the methods identified 46 vertebrate taxa from 30 families: 23 birds, 13 mammals and 10 fish. Seven taxa appeared through both methods, three through bone shape alone and 36 only through DNA. This is the same principle that has allowed ancient genetic traces to revise younger fossil histories, including ScienceBlog’s report on mammoths and horses detected in Yukon soil.
The species list joined habitats that are rarely preserved together. Polar bear, ringed seal, bearded seal, walrus and bowhead whale pointed toward sea ice. Blue whale and harbour porpoise added open-water animals. Atlantic cod, haddock and golden redfish represented the cold sea, while Arctic grayling and other freshwater fish required accessible rivers or lakes. On land were Arctic fox, reindeer, hare, tundra vole and collared lemming.
The lemming was new to Scandinavia, not extinct everywhere
The collared lemming, Dicrostonyx torquatus, is the finding most easily compressed into a misleading phrase. The fossil was the first evidence that the species once reached Scandinavia, and the animal is now absent from Fennoscandia. But the Mammal Diversity Database lists the species as living in Russia and classed as Least Concern.
Its range has nevertheless contracted dramatically. Genetic work on Eurasian collared lemmings indicates an eastward retreat from western Europe around the end of the last glacial period. The Norwegian specimen also fell on a mitochondrial branch that has no living descendants. In other words, the species survived, but the population lineage represented in the cave did not.
The same pattern appeared in the sequenced polar bear and Arctic fox material. Their mitochondrial genomes belonged to extinct lineages too. Species names can hide that turnover: a polar bear living today is not evidence that every ancient polar bear population made it through repeated changes in ice and habitat.
The cave reconstructs a coastline with no modern twin
The mixture suggests open Arctic or subarctic tundra beside seasonal or perennial sea ice, with nearby freshwater and at least some pine. Reindeer needed enough ice-free ground to migrate. Freshwater fish needed connected water bodies. Sea-ice mammals and harbour porpoise point in different directions, which makes a seasonally changing coast more plausible than one locked permanently under ice.
The bones probably did not represent animals that all lived inside the cave. The team found no burning, cut marks or other evidence of human accumulation. Predators may have brought some prey inside, and marine mammal fragments may have come from scavenged carcasses. Walrus and whale material was identified only through DNA, possibly because lipid-rich marine bones preserve poorly and had become too fragmented for visual identification.
This was also not the famous mammoth steppe of woolly mammoths, horses and rhinoceroses. It was a coastal mosaic. The Bournemouth University announcement called it a vanished Arctic world, but its scientific interest lies in how many familiar species occupied an unfamiliar combination of habitats.
An ancient cold shift is not a direct model of today’s warming
When colder, fully glaciated conditions returned, northern Norway apparently did not remain a refuge for these populations. The authors interpret the extinct mitochondrial lineages as evidence that the animals could not track suitable habitat locally through the later glacial advances. Some species later recolonised the region from elsewhere; others no longer occur there.
That is relevant to climate vulnerability, but it is not a controlled forecast of the modern Arctic. The ancient transition moved toward deeper cold and expanding ice. Today’s pressure is rapid warming, shrinking sea ice and landscapes fragmented by human activity. The direction, speed and available routes differ.
What Arne Qvamgrotta demonstrates is more fundamental. A species can persist on a map while local populations, genetic diversity and entire ecological combinations disappear. The cave did not preserve a simple parade of Ice Age icons. It preserved the evidence needed to reconstruct who lived together, how land, fresh water and sea ice met, and how much biological history can be lost even when some familiar names survive.