Two studies published in Nature report that domestic dogs lived alongside Ice Age hunter-gatherers across western Eurasia as early as 15,800 years ago, pushing back the genetic confirmation of dog domestication by roughly 5,000 years. The research centers on ancient DNA extracted from canid remains found at sites in Somerset, England, and Central Anatolia, Turkey, placing dogs firmly in the Late Upper Palaeolithic period, thousands of years before the rise of agriculture. The findings reshape the timeline of one of the oldest partnerships in human history and raise fresh questions about where, how, and how many times wolves became dogs.
What is verified so far
The strongest evidence comes from a jawbone housed at the Natural History Museum in London. That specimen, recovered from Gough’s Cave in Somerset, dates to approximately 14,300 years ago and has been genetically identified as a domestic dog rather than a wild wolf. Genome-wide ancient DNA analysis separated this animal from its wolf relatives based on clear markers of domestication, making it one of the oldest confirmed dog genomes anywhere in the world.
A second set of canid remains from Pınarbaşı in Central Anatolia yielded genome-wide DNA dated to roughly 15,800 years ago. That specimen, too, was classified as a domestic dog, and its analysis was also accessible via a separate institutional login route for subscribers. Together, the two finds show that dogs were not confined to a single region during the Late Upper Palaeolithic but were spread across a wide geographic band stretching from Britain to Turkey.
The Late Upper Palaeolithic spans roughly 16,000 to 14,000 years ago, according to a summary from Oxford. That window places these dogs squarely in the Ice Age, when small bands of human hunters tracked game across tundra and steppe. Dog domestication, in other words, happened in a world without farms, villages, or grain stores. The relationship between humans and canines began not as a byproduct of settled life but as a partnership forged under extreme environmental pressure.
A companion study analyzed 216 ancient canid remains, of which 181 came from Palaeolithic and Mesolithic Europe, using a genome-wide capture approach to sort dog from wolf ancestry across a large sample. That dataset includes what the researchers describe as the oldest dog genomic data they recovered, from a specimen dated to approximately 14,200 years old. Additional early dog remains were also identified across Germany, according to a press release summarizing the findings, though detailed genomic sequences and dating reports for those specimens have not yet been made publicly available.
Dietary and isotope work was also conducted on these early dogs. Isotope analysis can reveal what an animal ate and, by extension, whether it shared food sources with nearby humans. The institutional summaries reference this work, but primary analytical results and raw data have not been released in full, limiting independent verification of specific dietary conclusions. Until those datasets are accessible, claims about whether Ice Age dogs relied mostly on human-provided food, scavenged carcasses, or hunted for themselves will remain provisional.
What remains uncertain
The biggest open question is not whether dogs existed 14,000 to 16,000 years ago but how they got there. The broader scientific debate over dog origins remains unresolved, and these new studies sit within a long-running argument about whether domestication happened once or multiple times, and in which part of the world.
Earlier research using ancient wolf genomes proposed what scientists call a “dual ancestry” model for dogs. That work, published in Nature, assessed scenarios in which dogs descend from multiple wolf populations rather than a single founding event followed by later mixing. The new findings do not settle this debate. They confirm that dogs were present across a vast stretch of Eurasia by the Late Upper Palaeolithic, but they do not pinpoint where the initial split from wolves occurred or whether it happened more than once.
A Nature News and Views commentary linking the two primary studies, led by Bergström et al. and Marsh et al., frames the results as progress toward answering these questions without claiming resolution. The commentary, available directly and also through an alternative access path, outlines what “oldest dog genomes” means in scientific context and highlights remaining questions, including the geographic origin of domestication and the number of independent wolf-to-dog transitions.
There is also a minor discrepancy in the reported dates that readers should be aware of. One study describes the Pınarbaşı canid as dating to approximately 15,800 years ago, while the companion study’s oldest dog genomic data comes from a specimen dated to roughly 14,200 years ago. These are different specimens from different sites, analyzed by overlapping but distinct research teams. The gap does not represent a contradiction so much as a reflection of the complexity of working with ancient remains across multiple excavations and dating methods. Still, the difference means that claims about “the oldest” dog genome depend on which dataset and which analytical framework is being referenced.
The implications of dual wolf ancestry for understanding human migration patterns also remain speculative. Some interpretive commentary in the Nature articles suggests links between specific dog lineages and human movements across Eurasia, but no direct primary evidence yet maps those connections with certainty. The Associated Press framed the two Nature studies within this broader debate, noting that the exact location and timing of domestication remain contested among researchers and that different teams emphasize different candidate regions based on their datasets.
Another uncertainty concerns how representative these early dogs were of the broader canid population. The animals from Gough’s Cave and Pınarbaşı were preserved well enough for genome-wide sequencing, but preservation is patchy across time and space. Many ancient canid bones never fossilized or have yet to be excavated, and others have yielded only partial DNA. As a result, the current record might overemphasize regions where preservation conditions are favorable or where archaeological work has been especially intensive. It is entirely possible that equally old or older dogs existed elsewhere but have not yet been identified genetically.
Finally, the behavioral side of domestication (how early dogs actually lived with humans) remains largely inferred. Archaeologists can sometimes infer close bonds from burial practices, such as dogs interred alongside humans, or from pathologies suggesting care for injured animals. For these Late Upper Palaeolithic dogs, however, the published summaries focus primarily on genetics and dating. Until more contextual details about their archaeological layers, associated artifacts, and any evidence of deliberate burial are available, reconstructions of day-to-day human–dog relationships at these sites will remain cautious.
How to read the evidence
Not all of the evidence behind these findings carries the same weight. The strongest claims rest on genome-wide ancient DNA, a technique that extracts and sequences genetic material preserved in bone over thousands of years. When researchers can recover enough DNA to compare a specimen against known wolf and dog genomes, the classification is relatively firm. The Gough’s Cave jawbone and the Pınarbaşı remains both cleared this bar, which is why they are described as genetically confirmed domestic dogs rather than “wolf-like” or ambiguous canids.
The second tier of evidence involves the large-scale companion study of 216 ancient canid remains. A genome-wide capture approach was applied to distinguish dog from wolf ancestry across this sample, but the quality and completeness of ancient DNA varies widely from specimen to specimen. Some of those 181 Palaeolithic and Mesolithic European samples likely yielded more fragmentary data than others. The study’s value lies in the breadth of its sample, which allows statistical patterns to emerge even when individual specimens are imperfect. But the certainty attached to any single classification within that dataset is not uniform, and individual outliers may be reinterpreted as more data become available.
The third and weakest tier involves interpretive and contextual claims. Statements about what domestication “means” for understanding human behavior, or about how dogs may have enhanced foraging efficiency, are analytical inferences rather than direct observations. They are reasonable hypotheses, but they depend on assumptions about Ice Age ecology and human-animal interaction that cannot be directly tested with DNA alone. Isotope analysis, which can reveal dietary overlap between dogs and humans, offers one way to test these ideas. The institutional summaries confirm that such work was conducted, but the detailed results are not yet publicly available for independent review, so readers should treat specific claims about diet or cooperative hunting as provisional rather than definitive.
One common assumption in popular coverage of dog domestication deserves scrutiny: the idea that finding old dog remains automatically tells us where domestication began. It does not. A 15,800-year-old dog in Turkey and a 14,300-year-old dog in England tell us that dogs were already widespread by those dates. The actual origin could be older and located somewhere else entirely, in a region where preservation conditions or excavation efforts have not yet produced comparable specimens. The absence of evidence from Central Asia or East Asia in these particular studies does not mean those regions played no role. It means the current genetic record is incomplete and biased toward regions with long archaeological traditions and well-studied cave sites.
Another point to keep in mind is that “domestic dog” in a genetic sense does not necessarily look like a modern pet. The genomes of these Ice Age animals cluster with dogs rather than wolves, but their bodies may have resembled robust, wolf-like canids rather than the diverse shapes seen today. Domestication is a gradual process, and the earliest dogs likely occupied an intermediate space: genetically distinct enough from wolves to be recognizable in DNA, but still behaviorally and morphologically close to their wild relatives. The new studies primarily address that genetic threshold, not the full suite of physical and behavioral traits that later came to define dogs in human societies.
For readers trying to assess the significance of these findings, the most reliable takeaway is also the simplest. Dogs were living with human hunters across a wide geographic range during the Ice Age, well before anyone planted a crop or built a permanent settlement. The genetic evidence for that claim is now stronger and older than it has ever been. The harder questions, about exactly when wolves first became dogs, where that happened, and whether it happened more than once, remain genuinely open. These studies narrow the possibilities without closing the case, and they provide a clearer framework for testing future discoveries against existing models of dog origins.
Why the timeline matters
Placing confirmed dogs in the Late Upper Palaeolithic has implications beyond canine history. It suggests that hunter-gatherer societies were capable of forming and sustaining complex, long-term relationships with other species long before agriculture. Rather than emerging as camp followers at early farming villages, dogs appear to have been woven into the social and economic fabric of mobile foraging groups. That, in turn, hints at sophisticated planning, cooperation, and emotional bonds between humans and animals in Ice Age Europe and western Asia.
This earlier timeline also reshapes how researchers think about the co-evolution of humans and dogs. If dogs were already established companions by 15,800 years ago, then many later developments, such as the spread of humans into new environments, shifts in hunting strategies, and the rise of symbolic practices involving animals, may need to be re-examined with dogs in mind. The presence of dogs could have influenced which prey species were targeted, how groups organized their movements, and how they perceived and represented the natural world.
At the same time, the new studies underscore the limits of what genetics alone can reveal. Ancient DNA can identify lineages, trace ancestry, and refine timelines, but it cannot fully capture the lived experiences of humans and dogs sharing Ice Age landscapes. For that, archaeologists will need more integrated evidence: detailed site reports, careful analysis of bone modifications, contextual information about hearths and living spaces, and eventually, broader comparative work across regions that are currently underrepresented in the genetic record.
What the research does establish is that the bond between humans and dogs is not a product of civilization. It predates it by thousands of years. The partnership was born in a world of glaciers, megafauna, and small, mobile bands of hunter-gatherers, and it has persisted through radical transformations in climate, technology, and human society. As new sites are excavated and more ancient genomes are sequenced, that story will almost certainly become more complex. But the core insight, that humans and dogs have been companions since the Ice Age, now rests on some of the firmest genetic evidence yet assembled.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
