When gold miners in Canada's Klondike region melted permafrost to reach buried gold, they exposed something more valuable to science than the ore itself: ancient ground-squirrel burrows sealed in ice for up to 700,000 years, packed with fossilized feces that turned out to be a continuous genetic archive of Ice Age life in North America. The DNA preserved in those coprolites, the technical term for fossilized droppings, is now the longest record of its kind pulled from a single animal group on the continent, according to a study published this week in Nature Communications (Murchie T.J. et al., 2026, Nat Commun 17, 4868; https://doi.org/10.1038/s41467-026-72977-6) and reported by Scientific American.
The finding is more than a curiosity about squirrel diets. It expands the reach of ancient DNA research into an animal that modern ecologists know mainly as a seed-eating hibernator. It documents a previously unknown lineage within the ground-squirrel genus Urocitellus, a group of 13 species, mostly distributed in northwestern North America and Asia, that includes the Arctic ground squirrel. And it contains what may be the oldest mammoth DNA recovered anywhere in North America.
The Pleistocene, the last great ice age, ran from roughly 2.6 million to 11,700 years ago. Perennially frozen ground, or permafrost, has acted as a deep freeze for biological material in places like the Klondike. When mining operations use high-pressure water jets to thaw frozen sediment and sift for gold, the meltwater cuts into burrows that have been entombed since before the last ice age ended. The burrows belonged to ancient ground squirrels that spent up to eight months a year in torpor, a hibernation-like state in which body temperature drops and metabolism slows. After emerging each spring, they defecated in their burrows, and those droppings were buried, frozen, and preserved.
A research team led by evolutionary geneticist Tyler Murchie at the Hakai Institute in Heriot Bay, Canada, extracted and sequenced DNA from the coprolites, which range in age from roughly 17,000 to 700,000 years, as reported in Scientific American. The sequences include genetic material from woolly mammoths, steppe bison, and large cats, alongside the squirrels' own DNA. The presence of megafauna DNA in squirrel droppings is indirect evidence of behavior: the squirrels appear to have eaten, or scavenged, the carcasses of those large animals. Evolutionary biologist Bryan McLean at the University of North Carolina at Greensboro told Scientific American that the animals emerged from torpor "desperate for protein," and the carcass scavenging is consistent with a post-hibernation landscape already littered with the dead of winter.
The dietary record is reconstructed using a DNA sequencing technique that matches short genetic sequences in a sample to a reference database of known species — a methodology the study's authors describe in detail in the Nature Communications paper. The result is a list of what passed through the squirrels' digestive systems, not a direct observation of behavior. The authors caution that the detection of large-mammal DNA does not by itself prove the squirrels were regular scavengers of mammoth, bison, or big-cat carcasses. It does, however, place these hibernators squarely inside the Pleistocene scavenging chain, a real ecological role rather than a metaphor.
Beyond diet, the coprolite DNA tracks the squirrels themselves across hundreds of thousands of years. Most striking is a previously unknown lineage within Urocitellus that does not match any of the genus's modern species. The lineage appears, persists through several climate shifts, and eventually disappears, leaving no obvious descendants in today's fauna. The mechanism by which the lineage went extinct is not yet clear from the genetic data alone, but the fact that the coprolite record captures the rise and fall of a distinct squirrel population is, in practical terms, a snapshot of evolution in motion.
The mammoth DNA recovered from the same samples may be the oldest yet sequenced from North America. The source material is described in the popular write-up as potentially the continent's oldest, a hedge that the underlying peer-reviewed paper qualifies carefully, and the Scientific American account preserves the same caveat. The candidate record would extend the known reach of mammoth genetics on the continent by tens of thousands of years if confirmed against existing permafrost-preserved mammoth remains.
For ancient DNA researchers, the larger lesson is methodological. Genetic material in permafrost is well known to survive in bones, teeth, hair, and plant macrofossils. A continuous archive preserved in the droppings of a small hibernating rodent is something different: a low-cost, high-resolution way to monitor an ecosystem across deep time. The Pleistocene of the Klondike, accessed inadvertently through gold mining, has become a baseline case study for what burrowing hibernators can leave behind.
The miners' jets are still running. Each thaw cycle exposes more burrows, and with them, more coprolites. The research team plans to extend the analysis to other Yukon sites and to other small mammals whose burrows intersect permafrost. The next set of questions is whether the unknown squirrel lineage left traces elsewhere, whether the mammoth DNA holds up against independent samples, and whether the same approach can recover plant and fungal DNA from the same burrows. The Pleistocene, in other words, is not finished giving up its secrets, and the archive is, in a literal sense, full of them.