The tomb had been used in two distinct phases. The first, running roughly from 3200 to 3100 BC, was a community burial in the fullest sense. The second, beginning around 2800 BC and continuing until about 2470 BC, was something else entirely. Between them: silence. A gap of several centuries when no one was buried there at all. The people who returned were not the descendants of those who left.

What happened in between is the question the paper is built around, and the answer, insofar as one exists, is multipart, contested at its edges, and deeply strange.

Read more

Read more

A new study in Scientific Reports1 by Sabine Gaudzinski-Windheuser, Lutz Kindler, Wil Roebroeks, and colleagues presents a detailed analysis of those fragments and makes a claim that has real geographic weight: this is the first documented evidence of Neanderthal turtle exploitation north of the Alps and Pyrenees, outside the Mediterranean basin where such evidence has been accumulating for decades.

The finding is not primarily about adding a species to a list. What makes the Neumark-Nord turtle assemblage interesting is the interpretive problem it creates. When Neanderthals at sites like Kebara in Israel or Gruta da Oliveira in Portugal collected tortoises intensively enough to affect local population sizes, the conventional explanation involves resource pressure—broad-spectrum foraging as a response to the scarcity or difficulty of obtaining larger prey. That logic cannot apply here. At Neumark-Nord, the large mammals were abundant, intensively exploited, and clearly the caloric foundation of Neanderthal subsistence at this location. A one-kilogram pond turtle in that context is not a substitute for anything. So why collect them at all?

Read more

A study published this month in PNAS Nexus1 brings that argument to its largest scale yet. Rozenn Colleter, Michael P. Richards, and colleagues analyzed stable carbon and nitrogen isotopes from 12,281 adult Europeans spanning 393 sites and roughly 10,000 years of history, from Mesolithic hunter-gatherers to the early modern period. Their central finding: men have been consistently overrepresented among high meat-consumers across virtually every period and culture in the dataset. Women dominate the low end. That pattern is not continuous or invariant, but it is persistent enough to constitute something structurally significant.

The methodological innovation here deserves attention before the findings do. Comparing isotope values across sites is notoriously difficult. Nitrogen isotope ratios in bone collagen track animal protein intake, but those ratios reflect local environmental baselines—the degree to which farmers used manure fertilizer, the presence of freshwater food chains, climate variation, even metabolic stress during pregnancy. A high δ¹⁵N value in a coastal medieval Danish site means something different than the same value from an inland Roman-period Italian one. Direct comparison produces noise rather than signal.

Read more

A stratified open-air site in the grasslands of KwaZulu-Natal has now disrupted that picture in a significant way.

Jojosi sits roughly 140 kilometers from the Indian Ocean coast in eastern South Africa, a landscape of eroded gullies cut by the Jojosi River through a dolerite foot slope. The terrain is geologically active: cut-and-fill cycles have been reshaping it since at least 600,000 years ago, alternately burying and exposing sediment. Since 2022, an interdisciplinary team led by Manuel Will at the University of Tübingen has been excavating five artefact lenses at three locations within these gullies. What they found, and what they have now published in Nature Communications,1 challenges the embedded procurement model directly.

Every single stone artefact recovered from the stratified assemblages at Jojosi — across all lenses, across all sites, totalling more than 20,000 lithic pieces from the excavations alone — is made from hornfels. Not dolerite, which outcrops immediately adjacent to the site. Not quartz or quartzite, available as river gravels in the Jojosi River. Only hornfels: a fine-grained metamorphic rock formed when argillaceous siltstone is baked by contact with hot igneous intrusions, yielding a grey to dark grey material with low knapping force requirements and excellent flake quality.

The site also contains large primary hornfels outcrops, roughly 500 meters from the excavated lenses, as well as angular blocks that have been washed down into the alluvial gravels and gully fills. What it does not contain — in the stratified deposits or in the extensive surface record covering approximately one square kilometer — is meaningful evidence of anything else. No retouched tools. No retouching flakes (which would indicate tool finishing on-site). Almost no end products at all: large blades and preferential flakes were the apparent goal of the knapping, but they are largely absent. Use-wear analysis of 40 non-cortical blanks found traces of use on exactly one piece. There is no substantial faunal assemblage, no evidence of hearths, no indication of the broad range of activities associated with residential occupation.

What the assemblages do contain is abundant cortical material, core preparation and rejuvenation flakes, and vast quantities of microdebris: at Jojosi 6 alone, two lenses yielded nearly 5,000 and 3,800 pieces of material under 5 millimeters respectively, a size distribution that matches experimental hornfels knapping workshops. Refitting work on 353 artefacts across all sites, producing 123 refit groups with spatial extents under 30 centimeters, confirmed that the lenses represent short, discrete knapping events with high stratigraphic integrity. Pieces fit back together into reduction sequences — decortification of large blocks, core preparation, blank production — that end with the blanks gone. Knappers came, reduced hornfels, took the products, and left.

Read more

What happened next is what a team of researchers spent years trying to reconstruct, and the answer they arrived1 at has pushed back the known record of hominin megafaunal butchery at Olduvai Gorge by roughly 300,000 years.

The site is called EAK — the Emiliano Aguirre Korongo — and it sits at the junction of Olduvai’s two gorges in Tanzania, sandwiched in time right at the Bed I/Bed II boundary, dated to approximately 1.78 million years ago. The partial remains of a single juvenile E. recki individual were excavated there beginning in 2022, after erosion from seasonal rains exposed bones that had been buried for nearly two million years. What emerged was a pelvis, both hindlimbs, portions of eight ribs, a skull positioned upside down with its tusks still projecting, and eighty stone tools scattered tightly among the bone. The tools were made mostly of quartz, mostly small flakes and their debris, the kind of expedient kit you’d expect from someone who came to work rather than to display.

The question — the one that has dogged every elephant site from East Africa to the Levant — is whether the tools and bones ended up together because hominins put them there, or because geology, carnivores, and coincidence assembled a convincing fake. That question is surprisingly hard to answer, and the researchers behind the EAK study put real methodological effort into taking it seriously.

Read more

Not gravel. Not sand. Shells. Between 70 and 90 percent of the material that makes up this 3,000-square-meter island is the discarded remains of edible shellfish: clams, cockles, gastropods, species that people eat. The rest is a sandy-clay matrix, with occasional fragments of plain pottery mixed through.

Patrick Nunn and colleagues first came across the site in January 2017 during a reconnaissance survey along the northern coast of Vanua Levu. They initially thought it was a promontory attached to the mainland. Detailed mapping in 2024 showed it was something else — a discrete island, surrounded by water and mangroves, rising perhaps 20 to 60 centimeters above mean high tide. The team returned twice that year, excavated four test pits, put down 20 narrow hand-auger cores, and sent ten Anadara shells for radiocarbon dating. Their findings are published in Geoarchaeology.1

The radiocarbon dates came back clustered. Tightly. The median age across all ten samples was 1,190 calibrated years before present, or roughly 760 CE. The full range runs from about 420 to 1040 CE, a span of six centuries. The clustering was unexpected enough that the team ran a second phase of dating specifically to check it.

Read more

Those bones are still around. A few of them are in storage at the Smithsonian, at the University of Wyoming, at the Denver Museum of Nature and Science. They have been catalogued, photographed, and largely filed away as ambiguous “gaming pieces” for decades. What they apparently are, according to a new study published in American Antiquity,1 is the oldest dice ever found.

The paper, by Robert J. Madden of Colorado State University, argues that Folsom-period hunter-gatherers at the Agate Basin site in Wyoming, the Lindenmeier site in Colorado, and Blackwater Draw in New Mexico were making and using two-sided bone dice — what historians of gaming call binary lots — no later than 12,000 years ago. That places the invention of dice not in Bronze Age Mesopotamia or the Indus Valley, where the earliest previously known examples date to roughly 3500 BC, but in North America, at the tail end of the Ice Age, among highly mobile bison hunters who left no cities, no writing, and no permanent settlements.

The gap between the two is more than 6,000 years.

Read more

What makes the site at Didé West 1 (DDW1) unusual is not just its age — though a start date around the 4th century BCE is early by any measure for sub-Saharan Africa — but its stubborn coherence. Over nearly 800 years, the smelting tradition practiced there barely changed. That kind of continuity is rare in the archaeological record, and it raises questions that go well beyond this single site.

Excavations in 2018 and 2022, carried out by an international team coordinated by the University of Geneva (UNIGE) in collaboration with the Institut Fondamental d’Afrique Noire (IFAN) in Dakar, exposed the remains of this iron-smelting workshop in exceptional condition: 35 circular clay-lined furnace bases, each roughly 30 centimeters deep. A semicircular arrangement of around thirty used tuyères surrounded the furnace area. Stratigraphic and spatial analyses show that the workshop’s footprint gradually shifted northward over time, a slow drift that gives the site its chronological structure. The whole complex sat beneath that massive slag heap. The paper reporting the findings appeared in African Archaeological Review.1

Read more

A new paper in Archaeology in Oceania1 doesn’t fully resolve that puzzle, but it adds to it in interesting ways. A team led by Paul Taçon of Griffith University, working in collaboration with Traditional Owners, has documented 14 previously undescribed paintings of Thylacinus cynocephalus and two of Sarcophilus harrisii at two sites in western Arnhem Land: Awunbarna (known to non-Aboriginal Australians as Mount Borradaile) and Injalak Hill, near Gunbalanya. The images span multiple art styles and potentially multiple millennia, and at least some of them may be less than a thousand years old.

That last point matters. The consensus extinction date for both species on mainland Australia sits around 3,000 years ago, probably driven by competition with dingoes arriving between 3,500 and 5,000 years BP, combined with hunting pressure and possibly climate change. White et al. (2018) placed the mainland extinctions for both species between 3,179 and 3,227 years BP. But a painting made in white kaolin pipe clay — which weathers quickly and rarely survives more than a few centuries on exposed sandstone — is a different kind of evidence. If the paintings are what the researchers think they are, someone in Arnhem Land was still putting these animals on rock long after the mainland extinction date.

Joey Nganjmirra identifies a thylacine. Credit: Andrea Jalandoni

Read more

This is what a genuine local adaptation looks like. Herman Pontzer, an evolutionary anthropologist at Duke University, uses examples like this one throughout his new book, Adaptable: How Your Unique Body Really Works and Why Our Biology Unites Us (Penguin Random House, 2025), to illustrate something he treats as central to what our species is. We are, above almost everything else, flexible.

“That’s why there’s 9 billion of us and not 9 billion of some other primate,” Pontzer told Live Science. Adaptability is the mechanism. It’s what allowed Homo sapiens to occupy every biome on the planet, adjusting through culture, technology, and accumulated biological change across generations. No other primate comes close.

The book is a tour of the human body, system by system, with particular attention to how environments have shaped what those systems do. Pontzer has spent years working with the Hadza of Tanzania, a contemporary hunter-gatherer population, and that research gives him a frame of reference for thinking about baseline human physiology that most biomedical research lacks. Working with diverse populations doesn’t just add data points; it changes the questions you know to ask.

What Local Adaptation Actually Requires

The Sama spleen story is clean, but Pontzer is precise about what makes it possible. Local adaptations are real and documented, but the conditions that produce them are narrow. A trait has to help individuals survive and reproduce in one specific place. Not everywhere, just there. It has to persist across enough generations for selection to accumulate. And the environment driving it has to be stable enough, and geographically bounded enough, that gene flow, the constant mixing of alleles through interbreeding between populations, doesn’t dilute the effect before it takes hold.

Most traits don’t survive those criteria.

Skin pigmentation does. The gradient of ultraviolet radiation between the equator and the poles is old and consistent, and so is the gradient of melanin production across human populations. Darker skin offers protection against UV damage; lighter skin permits greater vitamin D synthesis where UV is scarce. Both directions of the trade-off have been advantageous in their respective environments for long enough that selection has had time to work. High-altitude adaptations in Himalayan populations follow a similar structure: the mountains have been high throughout the entire span of human prehistory, and so have the selection pressures they impose.

Other proposed local adaptations collapse under the same scrutiny. In the 1990s, some researchers argued that Black Americans might carry alleles predisposing them to hypertension and heart disease, the implication being that some evolutionary pressure had shaped cardiac function differently in West African populations. Pontzer is skeptical, for reasons that follow directly from the mechanics of local adaptation. Having a heart that functions well is not a localized advantage. It is useful everywhere. Traits that are universally beneficial spread through gene flow. They don’t concentrate in populations.

The same logic applies to claims, still circulating, about population-level differences in cognitive ability having evolutionary roots. Intelligence has been selected for across the entire species, continuously, for as long as Homo sapiens has existed. There is no environment where diminished cognitive ability was adaptive. Any variants that enhance brain function would be expected to spread broadly, not cluster. The framework that makes the Sama spleen story coherent is precisely what makes these other claims incoherent.

A Body Built for Somewhere Else

Pontzer’s work with the Hadza feeds into a second argument in the book: our bodies were shaped in an environment radically different from the one most people now inhabit, and the gap between those environments is doing measurable damage.

Hunter-gatherers are physically active continuously, eating from wild food sources, exposed to a wide range of pathogens. This was the norm for Homo sapiens, and for the hominin lineages that preceded our species, for millions of years. The body we have is a product of that context. Move it into a climate-controlled house with a caloric surplus and minimal required movement, and the same physiology fine-tuned for one environment starts producing maladaptive outcomes in another: heart disease, metabolic disorders, allergies, conditions that appear to have been rare before the agricultural transition and are common now.

This is the evolutionary mismatch, and Pontzer is careful not to let it slide into nostalgia or primitivism. It is a mechanistic observation, not a moral one. The body is responding rationally to its conditions; the conditions just happen to be novel in ways that have outpaced any biological adjustment.

One of the more interesting extensions of this framework involves genetics and development. Pontzer describes the genome as setting a range of possible outcomes rather than a fixed destination. Your genes constrain what you can become but don’t determine it. The environment usually has the larger visible effect on which possibilities actually materialize. Epigenetics adds another layer: environmental stresses can alter how genes are expressed, switching them on or off in ways that persist for a lifetime. In mice, these changes have been shown to transmit to offspring. The environment a mother experiences can affect her children’s biology. In humans, the evidence for transgenerational epigenetic inheritance is suggestive but not yet settled. The studies required take decades to run, and the full picture isn’t in yet.

What Pontzer keeps returning to is that diversity is real but layered in ways that resist simple categorization. Knowing something about a person’s pigmentation tells you essentially nothing reliable about their cardiovascular risk, their cognitive profile, or most other things you might want to know. The systems are largely independent. They evolved under different pressures, respond to different environments, and vary along different axes. The assumption that populations sort cleanly into types, and that a trait in one domain clusters reliably with traits in others, is precisely the error that the correct application of local adaptation logic should prevent.

Charles Perreault, an evolutionary anthropologist at Arizona State University’s Institute of Human Origins, has tried to do something that sounds straightforward but turns out to be genuinely difficult: measure that strangeness in evolutionary terms. His new paper in Proceedings of the National Academy of Sciences1 asks what it would have taken for a typical mammalian lineage to achieve the kind of planetary footprint our species has. The answer puts the role of culture into relief in a way that slogans about human uniqueness never quite manage.

To reach the geographic and ecological range Homo sapiens holds today through biological evolution alone, Perreault’s modeling suggests a lineage would have needed roughly 88 million years of divergence, more than 2,200 separate species, and a nearly four-order-of-magnitude spread in body mass. Our species has been around for approximately 300,000 years. We are one species. Our body mass variation, across all living humans, is trivial by any comparative mammalian standard.

Something else is doing the work.

Read more

The first digs took place in 1869 and 1870, when antiquarian Anders Lorange sank a shaft from the summit and cut trenches toward the center. He reached bedrock. No chamber, no grave, no remains. A second round in 1939 and 1940 under Sigurd Grieg opened a larger area and confirmed the absence. Grieg eventually concluded it was probably a cenotaph, a monument to someone buried elsewhere. A subsequent reanalysis of fragmented cremated bone from the 1940 excavations briefly revived the burial hypothesis. The bone, it turned out, dated to somewhere between 1391 and 1130 BC, centuries before the mound’s construction. It had been redeposited there, mixed into the soil. It was not evidence of a burial.

Raknehaugen, rising some 15 meters above a fertile plain roughly 40 kilometers north of Oslo and stretching 77 meters across, is the largest prehistoric mound in all of Scandinavia. It was built around AD 551. For 150 years, the standard interpretation has been that it commemorated a powerful individual. The mound, the thinking went, was a material expression of elite authority: the bigger the mound, the greater the person inside.

But Raknehaugen has been humoring this interpretation without supporting it.

A new study by Lars Gustavsen of the Norwegian Institute for Cultural Heritage Research, published in the European Journal of Archaeology,1 proposes a different answer. Gustavsen argues that Raknehaugen was a communal ritual response to a catastrophic landslide, not a burial monument at all. The evidence, once you look at it sideways rather than straight down through excavation shafts, turns out to have been hiding in the landscape the whole time.

Read more

The uppermost mounds date mostly to the Ptolemaic and Roman periods. Walk the surface of Kom C, the southern settlement mound, and what you encounter is Roman-period material pushed up by the landscape. Beneath that is the Saite period. Below that, the Third Intermediate Period. Deeper still, the Old Kingdom, the Early Dynastic, and eventually the Predynastic. Each layer obscures the one below it, and the site sits in a floodplain where groundwater complicates any attempt to excavate the earliest levels.

This is the challenge that a team from Kafrelsheikh University, Egypt’s National Authority for Remote Sensing and Space Sciences, and Durham University recently took on at Kom C. Their approach1 combined two tools: a radar satellite and electrical resistivity tomography. The goal was not to dig blindly but to know, before breaking ground, what was there and at what depth.

Read more

The find became world-famous anyway. Lehringen entered the literature as the clearest possible case of Neanderthal elephant hunting: a complete wooden weapon, the first ever found from a Middle Paleolithic context, lying in anatomical association with the largest land mammal known to have lived in Pleistocene Europe. The picture was compelling. Then the skeptics arrived.

The problems were real. The spear’s tip showed use-wear traces that left open the possibility it functioned as a multi-purpose tool, perhaps a digging implement or a snow probe rather than a hunting weapon. More troublingly, some researchers argued the spear might have washed into proximity with the carcass at the margins of the paleo-lake, a coincidental association rather than a killing blow. Without proper documentation, there was no way to definitively rule this out. For decades the site existed in an interpretive limbo: famous, referenced constantly, but not quite settled.

A study published this year in Scientific Reports1 by Ivo Verheijen, Gianpiero Di Maida, Gabriele Russo, and Thomas Terberger represents the first systematic zooarchaeological analysis of the Lehringen faunal assemblage. Their findings do not simply shore up the hunting hypothesis. They reframe the site altogether.

What the bones say

The elephant at Lehringen was a male, probably around 30 years old, based on molar wear. Not yet fully grown, but already large. His long bone epiphyses were still unfused, consistent with that age estimate. His death from natural causes is unlikely: the analysis found only minor pathologies to the vertebral column, and an animal in the prime of life at this size would not simply die at a lakeshore and then happen to have a spear fall beside him.

The key evidence is on the ribs. Seven ribs or rib fragments carry cut marks. Most of these are on the lateral surface, perpendicular or diagonal to the long axis of the bone, consistent with filleting meat from the rump. That alone would support butchery, not hunting. But one rib fragment carries something more diagnostic: a series of cut marks on its internal face, the surface that faces inward toward the chest cavity. Cut marks on the internal face of a rib have a specific meaning. To make them, someone had to be working from inside the thorax. The carcass had to be fresh. The organs were the target.

This matters because access to the internal organs requires primary access to the carcass, before carnivores have opened it, before any substantial decomposition. The cut marks on the inner rib surface indicate the elephant was eviscerated while still in a state that made the organs worth recovering. The researchers note that carnivore gnawing is present on the distal ends of ribs and the dorsal spines of vertebrae, but it is limited. The bone surface shows no significant weathering. The picture is not of a carcass opportunistically encountered and partially scavenged. It is of a fresh animal, butchered by hominins who got there first.

The lithic assemblage, 25 Baltic flint flakes recovered near the elephant’s skull, reinforces this. Use-wear analysis on some of these pieces is consistent with defleshing activity. No cores were recovered, which probably means the knapping happened elsewhere. These were working tools, brought to the carcass.

No direct evidence of spear impact was found on the elephant’s skeleton. No wound channel, no bone lesion attributable to a thrown or thrust weapon. The absence is not definitive: soft tissue injuries leave no skeletal trace, and an animal struck in the flank or neck might show nothing on the bones that survived. The spear between the ribs remains the most literal association of weapon and prey in the Paleolithic record, and the fresh-carcass evidence, the prime-age male, the lithics with butchery wear, and the lack of any plausible alternative explanation for the elephant’s death all point in the same direction.

A wider spectrum

The elephant is not the only story at Lehringen. The site preserves faunal remains from multiple stratigraphic layers spanning different phases of the Eemian interglacial. What the new analysis makes clear is that Neanderthals were not visiting this lakeshore once for a single spectacular kill. They were coming back, exploiting whatever was available.

Aurochs (Bos primigenius) remains represent at least three subadult individuals, recovered from the basal peat layer beneath the marl that contains the elephant. The remains of the most complete individual include cut marks on the lateral ramus of a mandible and on a lumbar vertebra, indicating defleshing. Carnivore gnawing, probably from wolf, is also present on the aurochs bones, suggesting that other predators were working these carcasses too, though whether before or after Neanderthal butchery cannot be determined.

A single brown bear (Ursus cf. arctos) is represented by two bones: a rib fragment and a distal femur. Both show anthropogenic modification. The rib carries cut marks on its external surface, consistent with filleting. The femur shows cut marks from filleting on both faces, and also impact marks with cone fractures on the shaft. Someone fractured that femur to access the marrow. Brown bear femora contain substantial marrow fat, particularly relevant in late summer and autumn when bears are at maximum fat deposition. The researchers note that bears, beavers, and elephants share a common dietary appeal: high fat content. This may not be coincidental.

The beaver (Castor fiber) evidence is the most anatomically detailed. Pelvis fragments, skull elements, and mandibular pieces all carry traces of human activity. Cut marks on the ilium suggest disarticulation at the hip joint. Marks on the maxilla, near the zygomatic arch, are positioned to indicate severance of the masticatory musculature, perhaps to detach the mandible. Marks on the lateral surface of a mandible fragment are consistent with skinning. Beaver fur is dense and waterproof. That the Lehringen Neanderthals were skinning beavers, not just butchering them for meat, fits with documented patterns at other Middle Paleolithic sites including Krapina in Croatia and Taubach in Germany, and goes back even earlier at Bilzingsleben, where both the European beaver and the giant beaver (Trogontherium cuvieri) were being processed with stone tools.

The breadth of prey at a single open-air lakeshore site is what stands out. Palaeoloxodon antiquus, a megaherbivore weighing several tons. Bos primigenius, large and dangerous. Ursus cf. arctos, a bear. Castor fiber, a semi-aquatic rodent valued for both its flesh and its pelt. The assemblage also preserves wels catfish (Silurus glanis), pike (Esox lucius), pond turtles, herons, cormorants, deer of multiple species, rhino, wolf. Not all of these were prey. But the lake environment clearly concentrated resources in ways that Neanderthals recognized and returned to.

The researchers interpret this as opportunistic procurement rather than specialized hunting. The contrast is with sites like Mauran and La Borde in France or Salzgitter-Lebenstedt in Germany, where large numbers of a single prey species point to targeted, probably organized drives or ambushes. At Lehringen, the diversity of exploited species and the absence of large numbers of any one taxon suggest a different mode: taking what the landscape offered, across multiple visits, from the full range of what the Eemian interglacial around a paleo-lake could provide.

The spear itself

The Lehringen thrusting spear deserves its own moment. Made from yew (Taxus sp.), from the trunk rather than a branch, it required the removal of up to 39 knots and side branches. That is not a weapon thrown together at a kill site. Yew was not a random choice either: the wood is flexible and strong, properties that matter in a thrusting weapon that needs to absorb impact without shattering. The time investment in its production implies forward planning and a clear conception of what the tool was for.

The spear shows extensive use-wear, suggesting it had been used repeatedly before ending up at Lehringen. Its current curved shape is a result of post-depositional deformation, probably from the overlying weight of the elephant and sediment. Some researchers suggested that curvature and use-wear indicated it might have served other functions, including digging, and this remains technically possible. But a digging stick is not what you build from a carefully trimmed yew trunk with a hardened point and 39 removed branches.

The spear was recovered in seven pieces and is currently broken into eleven, all of which fit together. It is held in Hanover, preserved in beeswax, and remains one of the most extraordinary objects from the Paleolithic record anywhere in the world.

What Lehringen offers now, with a proper zooarchaeological foundation under it, is a site that actually supports what the 1948 discovery seemed to show. Neanderthals hunted a straight-tusked elephant. They butchered it fresh, opening the chest cavity to reach the organs. They made the tools to do it, brought them to the lakeshore, and left traces on bones that have lasted 125 millennia in remarkably good condition. They also came back, or never left, hunting aurochs from the peat below the elephant and processing bear and beaver on other occasions. The lake was a resource concentration point, and they knew it.

Further Reading

• Gaudzinski-Windheuser, S., Kindler, L., MacDonalds, K., & Roebroeks, W. (2023). Hunting and processing of straight-tusked elephants 125,000 years ago: Implications for Neanderthal behaviour. Science Advances, 9, eadd8186.

• Gaudzinski-Windheuser, S., Kindler, L., & Roebroeks, W. (2023). Widespread evidence for elephant exploitation by Last Interglacial Neanderthals on the North European plain. Proceedings of the National Academy of Sciences, 120(50), e2309427120.

• Thieme, H., & Veil, S. (1985). Neue Untersuchungen zum eemzeitlichen Elefanten-Jagdplatz Lehringen, Ldkr. Verden. Die Kunde, N. F., 36, 11–58.

• Gaudzinski-Windheuser, S., Kindler, L., & Roebroeks, W. (2023). Beaver exploitation, 400,000 years ago, testifies to prey choice diversity of Middle Pleistocene hominins. Scientific Reports, 13, 19766.

• Kindler, L., et al. (2025). Large-scale processing of within-bone nutrients by Neanderthals, 125,000 years ago. Science Advances, 11, eadv1257.

• Schoch, W. H. (2014). Holzanatomische Nachuntersuchungen an der eemzeitlichen Holzlanze von Lehringen, Ldkr. Verden. NNU, 83, 19–29.

Rujm el-Hiri, as the site is known in Arabic (Hebrew speakers call it Rogem Hiri), has been excavated, surveyed, debated, and nicknamed “the Israeli Stonehenge” by people who apparently found the Rephaim, the biblical race of giants traditionally associated with the region, a serviceable explanation for how something this large ended up in the middle of nowhere. Its date is still genuinely uncertain: proposed construction periods range from the Chalcolithic through the Early Bronze Age to the Late Bronze Age, depending on which finds from which part of the site a given scholar weights most heavily. Excavations have yielded little that resolves the question definitively.

The function has fared no better. It has been called a burial monument, a ceremonial gathering ground for dispersed pastoral tribes, a defensive installation, and — most influentially in the popular imagination — an astronomical observatory, with the northeast entrance thought to align with the summer solstice sunrise. A 1998 study by Aveni and Mizrachi set out the geometry in detail and became widely cited. The astronomical interpretation proved sticky, partly because it is elegant, partly because the site offers no better explanation.

What all these interpretations share, beneath their differences, is a common premise: Rujm el-Hiri is unique. No comparable structure had been found within its vicinity. That uniqueness was doing a lot of interpretive work. If something is one of a kind, it is easier to imagine it as a special-purpose monument rather than one instance of a recurring practice. The absence of parallels made it easier to project singular meaning onto it.

A new study published in March 2026 in PLOS One,1 led by Michal Birkenfeld of Ben-Gurion University of the Negev and co-authored with Olga Khabarova, Lev V. Eppelbaum, and Uri Berger, has demolished that premise. Using high-resolution satellite imagery collected over a twenty-year span, combined with geophysical modelling and spatial analysis, the team identified 28 previously undocumented large circular stone structures within a 25-kilometer radius of Rujm el-Hiri. Only two examples had been recorded before in any survey. The others had simply never been seen.

Read more

That uncertainty has long been a problem in Japanese Sue ware studies, and it turns out to be exactly the right kind of problem for a deep learning model to get its hands on.

A team led by Wataru Tatsuda and Hayata Inoue recently published results1 from training a 3D point cloud classifier on Sue ware from the Sanage kiln in Aichi Prefecture, one of the dominant production centers for this pottery tradition during the eighth through mid-ninth centuries. The model achieved an overall macro F1-score of 0.9320 across five vessel types. That number is worth pausing on. The hard part wasn’t building a classifier that worked well. The hard part was understanding what it struggled with, and why that struggle is archaeologically interesting.

Sue ware itself is a natural candidate for this kind of work. It’s an unglazed stoneware, gray to brown-gray, wheel-thrown and tunnel-kiln fired, standardized in production to a degree unusual among ancient ceramics. Researchers have attributed some of that standardization to direct state involvement in the manufacturing process. The shapes are consistent enough to support typology, yet variable enough to generate real taxonomic disputes. Among the everyday tableware types from Sanage, two in particular have caused persistent headaches: the Dish Body and the Bowl.

The distinction, in principle, is straightforward. Dish-type vessels have steep inner walls and flat bases. Bowl-type vessels have gently curving walls and rounded bases. In practice, the eighth- to mid-ninth-century material from Sanage includes a substantial number of pieces that combine features of both, and expert archaeologists labeling the same sherds don’t always agree on which category applies. This isn’t sloppiness. It reflects something real about the material record.

Read more

They wrote almost nothing about the homes their converts lived in.

Those homes still exist, or what’s left of them does. Across four islands, 69 ruined stone cottages survive in varying states of collapse, their walls made from blocks of coral cut out of the surrounding reef. Known in Mangarevan as ‘are po’atu, they account for more than half of all the colonial-era structures archaeologists have recorded in the islands. The missionaries taught local people to build this way — Polynesian converts learning European masonry techniques, cutting Acropora branch coral from near-shore reefs and from beach rock formations on the motu, the narrow coral islets that fringe the lagoon. The technique spread. The construction materials were local. The buildings were everywhere. And European sources, predictably, were mostly silent about who built them and when.

The problem with recovering that history is partly chronological. Radiocarbon dating, the default method for establishing age in archaeological contexts, becomes unreliable for organic materials less than about 400 years old. The Mangarevan cottages were built mostly in the 1830s to 1860s. Radiocarbon can’t resolve that timeframe with useful precision. Dating by artifact typology is possible — imported ceramics and glassware have known production date ranges — but this requires excavation, adds interpretive noise, and still only narrows things down to multi-decade spans.

Coral is a different material. A new study by James Flexner of the University of Sydney, published in Antiquity,1 demonstrates that uranium-thorium (U-Th) dating can be applied directly to the coral blocks in the walls of these buildings, producing construction dates accurate to within a few years.

How the Method Works, and Where It Gets Complicated

Coral skeletons incorporate uranium from seawater as they grow. Once the coral dies, uranium slowly decays to thorium. Because fresh coral contains essentially no thorium at death, the ratio of uranium to thorium in a sample is a measure of time elapsed since death. For the past 500 years or so, lower concentrations of decay products make this technically demanding, but it works if samples are carefully selected and processed under ultra-clean conditions. Flexner had 10 branch coral samples analyzed at the University of Queensland Radiogenic Isotope Facility.

The samples came from unweathered or minimally weathered branch corals embedded in Acropora sp. limestone blocks — either pulled from walls still standing or recovered from blocks that had clearly fallen from adjacent ruins. This is important: heavily weathered coral becomes chemically compromised, and a few of the samples in this study illustrate what that looks like. The watch tower at Mata-Kuiti point returned a date predating European contact with Mangareva by half a century. The sample was notably weathered. The boys’ school at Aukena — the Collège d’Anaotiki, whose construction dates are known from missionary documents as 1853 to 1858 CE — returned 1831 ± 2 CE from a more weathered sample where better material wasn’t accessible.

So sample quality matters. But the more interesting complication lies in what the dates actually measure. U-Th dating on coral tells you when the coral died, not when it was placed in a wall. There can be a gap. Coral harvested from the seaward edge of a living reef would have died shortly before it was cut and used. But builders also drew material from the shoreward limestone, where coral might have been dead for years or decades before it was quarried. Some blocks may have been reused from older structures — the archaeological literature on Mangareva has long suggested that pre-contact sacred sites called marae were cannibalized for building material during the missionary construction period.

Flexner’s team treats this as analogous to the “old wood” problem in radiocarbon dating of timber buildings: the tree whose ring you’re dating may have died long before it was incorporated into a structure. The U-Th date is best understood as a terminus post quem — a “no earlier than” marker — rather than a direct construction date. The building can’t be older than the coral in its walls; it can only be the same age or younger.

With those caveats in place, the dates from the seven undated Polynesian cottages are coherent and informative. Most cluster in the first decade or so of missionary presence: AKH-7 at 1834 ± 3, AKH-10 at 1840 ± 3, AKH-1 at 1841 ± 2, AKH-35 at 1844 ± 2, AKH-20 at 1846 ± 2. One outlier, AKH-11, returned a pre-contact date of 1779 ± 2, which Flexner interprets as probable reuse of coral from a pre-European structure — possibly a marae. The gap between that date and the known period of cottage construction is decades long, but not centuries. No sample showed the kind of age gap that would suggest systematic large-scale looting of ancient reef formations. The reuse, where it occurred, was modest.

The Pit Beneath the House

The most vivid result in the study involves a house on Akamaru called AKH-20 and a pit feature found beneath its floor.

When excavators opened a test unit in the southwest corner of the building, they found a pit — PN-318 — filled with an unusually high concentration of material: bone fragments, shell, glass, and iron artifacts. Food and drink, including what appears to be alcohol-related debris. Compared with the other houses sampled on the island, this pit was exceptional for the density and variety of its contents.

The initial hypothesis was that this represented the waste from a single feasting event, perhaps connected to the construction of the house itself. A garbage pit inside a tropical household would be impractical, which made it unlikely that the material accumulated over the course of normal daily life. More plausible was a one-time deposit, sealed beneath the foundation when the building went up.

The two U-Th dates support this interpretation. A coral block from the wall of AKH-20 returned 1846 ± 2 CE. A branch coral recovered from inside pit PN-318 returned 1848 ± 4 CE. The dates overlap within their error ranges. Flexner suggests that the coral from the pit probably fell off during the trimming of blocks ahead of construction — builders cutting and shaping coral on site, trimmings falling into the pit, the pit then sealed as the walls went up.

If that reading is right, the dates bracket a moment: a feast, a construction event, and a household coming into existence. The feast debris sealed beneath the floor of a Polynesian Catholic home in the 1840s. None of that appears in any missionary document.

The objects inside the walls of these buildings add to the picture. Glassware, cooking pots, and ceramics found in excavated contexts indicate how families organized their domestic lives under the mission system — how meals were prepared, how interior space was structured, how religious practice shaped household routines. The architecture and its contents together tell a story about what it meant to be a Polynesian convert in the mid-nineteenth century: a life partly reshaped by European Catholicism and partly continuous with older ways of being that the documentary record treats as invisible.

What Flexner’s study opens up is the ability to put those buildings in sequence, to understand which came first and how quickly the construction spread across the islands. The boys’ school on Aukena was a control site with known dates; the Polynesian cottages had none. Now several of them do. Applied more broadly across the 69 surviving structures, U-Th dating could produce a construction timeline detailed enough to ask new questions about how the mission built its community of converts — which households appeared earliest, which came later, whether there are spatial or social patterns embedded in the sequence.

The same technique could work well beyond Mangareva. Coral limestone buildings survive across the tropical Pacific, across the Caribbean, along the East African coast. Much of that architecture was built by people who left no written record, documented by Europeans who saw no reason to write about the domestic lives of colonized populations. In those places too, the coral in the walls keeps a chemical record of time. It just needed a method capable of reading it.

Further Reading

• Kirch, P.V. et al. (2021). Coordinated ¹⁴C and ²³⁰Th dating of Kitchen Cave rockshelter, Gambier (Mangareva) Islands, French Polynesia. Journal of Archaeological Science: Reports 35. https://doi.org/10.1016/j.jasrep.2020.102724

• Kirch, P.V. & Sharp, W. (2005). Coral ²³⁰Th dating of the imposition of a ritual control hierarchy in precontact Hawaii. Science 307: 102–104. https://doi.org/10.1126/science.1105432

• Sharp, W.D. et al. (2010). Rapid evolution of ritual architecture in Central Polynesia indicated by precise ²³⁰Th/U coral dating. Proceedings of the National Academy of Sciences USA 107: 13234–39. https://doi.org/10.1073/pnas.1005063107

• Schiffer, M.B. (1986). Radiocarbon dating and the “old wood” problem: the case of the Hohokam chronology. Journal of Archaeological Science 13: 13–30. https://doi.org/10.1016/0305-4403(86)90024-5

• Emory, K.P. (1939). Archaeology of Mangareva and neighbouring atolls. Honolulu: B.P. Bishop Museum.

• Laval, H. (1968). Mémoires pour servir à l’histoire de Mangareva: ère chrétienne, 1834–1871. Paris: Musée de l’Homme.

What happened next, genetically speaking, is what a new study published in Nature Communications1 sets out to reconstruct.

Researchers from the Centre for Palaeogenetics at Stockholm University analyzed ancient DNA recovered from 13 individuals buried in a section of the Maqbara of Madina Yabisa, the main urban Muslim cemetery of medieval Ibiza, discovered during construction work at 33 Bartomeu Vicent Ramon Street in Ibiza town. The burials date to 950–1150 CE. Simple earth pits, bodies placed on the right side facing southeast toward Mecca, no grave goods — with one exception, a burial containing two silver rings. These were people interred according to Islamic law, in a functioning urban cemetery on an island that had gone from empty to organized in a single generation.

From 30 sampled individuals, 13 yielded enough DNA for population genomic and metagenomic analyses. That’s a small number. What they show is not.

Who Was Buried Here

The genetic picture of these 13 individuals is anything but homogeneous. Principal component analysis places them across a wide swath of the ancestral landscape: two cluster within European population space, one sits close to North African populations, eight occupy positions somewhere between European, Middle Eastern, and North African, and two sit squarely in Sub-Saharan African genetic space.

The majority carry mixtures of Iberian and North African ancestry, consistent with the known demographic history of al-Andalus after the Islamic conquest of Iberia in 711 CE, when Imazighen (Amazigh, often called Berbers in older literature) formed the largest group of new arrivals. Y-chromosome haplogroups E1b1b1b1a1 and E1b1b1a1a1c2 — both common in North African Amazigh populations — appear in six of the nine males. One individual, s.107, clusters so closely with pre-European contact Canary Islanders and Moroccan Imazighen that the team interprets him as an Amazigh individual whose small European ancestry component likely reflects the ancient European-related ancestry already present in pre-Islamic North African populations, not recent admixture with local Iberians.

Two individuals, s.157 and s.313, are genetic outliers in the other direction: minimal North African ancestry, genomes that look much like pre-Islamic Iberians. Both are buried in a Muslim cemetery. The team reads them as muladíes — muwalladûn in Arabic, recently Islamized local Iberians who retained the genetic signature of their ancestry while adopting a new religious identity. The observation has a larger implication: on Ibiza in the 11th century, cultural and religious transformation was apparently decoupled from genetic admixture. People converted to Islam faster than they mixed.

Two People from Opposite Ends of the Sahel

The two Sub-Saharan individuals are the most striking members of this group.

Individual s.117 is male. His mitochondrial haplogroup is L3e1c, his Y-chromosome haplogroup E1b1a1a1a — both found primarily in Sub-Saharan African populations. When the team compared his genome against a comprehensive panel of modern African populations including groups from across Chad, Sudan, the Sahel, West Africa, and East Africa, he aligned most closely with present-day populations from southern Chad: the Sara and the Laal, speakers of a Nilo-Saharan language and an endangered language isolate, respectively.

Individual s.197 also carries Sub-Saharan ancestry — mitochondrial haplogroup L3b2 — but his genome tells a different story. He clusters with Gambian populations and the Senegal Bedik, groups from the Senegambia region of West Africa.

These two men, buried in the same small cemetery on a Mediterranean island, came from regions roughly 4,000 kilometers apart, connected by the trans-Saharan networks that Arabic sources describe in some detail. Historical records from Ibn ʿIdhārī, al-Bakrī, and the al-Bayān al-Mughrib tradition document the northward movement of enslaved individuals from the Lake Chad basin via the Kawar and Fezzan oases — through present-day Niger and Libya — toward North Africa and ultimately Iberia. Those same sources describe a military force of around 4,000 cavalry from the kingdom of Takrūr, in the middle-lower Senegal Valley, crossing into Iberia with Yūsuf b. Tāshfīn at the Battle of Zallaqa in 1086 CE. Takrūr had converted to Islam around 1030 CE, and trans-Saharan networks subsequently supplied the Almoravid armies with both captives and voluntary soldiers.

The radiocarbon dates of s.117 and s.197, corrected for marine reservoir effects given Ibiza’s island context, place both individuals after 1115 CE — in the second demographic pulse that reached the Balearics following the Almoravid conquest of Mallorca in 1115–1116 CE. The genomic evidence and the historical record align on the same point: these were people moved by the Almoravid military and slave systems, arriving on Ibiza as part of a larger resettlement and garrisoning process.

The coexistence of Chadian and Senegambian ancestry in a single small cemetery shows something the historical texts suggest but rarely make vivid: the al-Andalus of the 11th and 12th centuries was not drawing from one part of sub-Saharan Africa. It was drawing from both the western and central Sahel simultaneously, through distinct but overlapping networks.

Timing the Mixture

Beyond identifying ancestral origins, the team used haplotype-based local ancestry analysis to estimate when North African admixture actually occurred in the ancestors of these individuals.

The method works by measuring the length of uninterrupted chromosomal segments of North African ancestry. When two populations mix, the resulting chromosomes carry long stretches of one ancestry type unbroken by recombination. Each subsequent generation, recombination shuffles those stretches shorter. The rate of that decay maps onto generational time.

The population-level estimate, combining seven individuals radiocarbon-dated between 1073 and 1094 CE, places the main admixture event approximately 7.84 generations before 1080 CE. Using a generation time of 26.9 years, that translates to roughly 869 CE — predating Ibiza’s conquest in 902 CE by about 33 years, and likely reflecting admixture that began in mainland al-Andalus before or during the initial colonization. Individual-level estimates range from 2.49 to 7.81 generations, pointing to ongoing gene flow rather than a single founding event.

One individual, s.157 — one of the two with predominantly European ancestry — shows an admixture date going back approximately 16 generations, to around 519 CE. This is pre-Islamic. The team interprets this as an ancient, low-level North African genetic contribution predating the Muslim conquest, probably connected to the demographic transformations of the late Roman period. S.157 may represent a family line that had a small North African contribution centuries earlier, absorbed into the local Iberian population long before Ibiza became part of al-Andalus.

Two individuals, s.131 and s.315, show elevated runs of homozygosity consistent with consanguinity, with s.131’s pattern suggesting first-cousin parental relatedness. This kind of endogamy was documented among Imazighen communities and in other small medieval Iberian populations. S.131’s admixture timing of 3.47 generations suggests that both parents were already part of the admixed local gene pool of Ibiza — a community that had been mixing for generations and was now marrying within itself.

The Leprosy Case

Individual s.313 is one of the two with minimal North African ancestry, almost certainly a muladí. His bones show no diagnostic skeletal signs of leprosy — some facial elements are missing, which makes a definitive osteological assessment impossible, but what survives looks unremarkable.

His DNA tells a different story.

Metagenomic screening of s.313’s shotgun sequencing data detected Mycobacterium leprae, the bacterium responsible for leprosy. Target enrichment with a custom capture panel increased the read yield from roughly 29,000 reads to 128,800, enabling a mean genome coverage of 3.75x — enough for phylogenetic placement.

The M. leprae genome from s.313 belongs to genotype 2F, a clade containing seven ancient genomes from across medieval Europe, dated between approximately 650 and 1250 CE. The clade stretches from Hospital of Sant Llàtzer/Santa Margarida in Barcelona to Sigtuna, Sweden. Maximum parsimony analysis tentatively groups s.313 with an individual from medieval Denmark (Jorgen749, 1223–1279 CE), though this specific relationship was not strongly supported in the maximum likelihood reconstruction.

The distribution of genotype 2F across medieval Europe illustrates something that individual burial evidence tends to obscure: disease was moving across the continent along the same networks that moved people, goods, and armies. Ibiza was not isolated from that. A bacterium recovered from a Muslim muladí in the western Mediterranean shares a lineage with individuals buried in Sweden.

S.313’s burial was indistinguishable from others in the cemetery. No sign of marginalization, no separation from the community. This is consistent with both Islamic theological and legal frameworks of the period, which placed considerable emphasis on the obligation to care for the sick, and with a broader pattern, also documented in contemporary Christian cemeteries, of including individuals with leprosy in standard community burials rather than excluding them. Whether s.313 had visible disease at the time of death, or died before symptoms appeared, or had a subclinical infection that never progressed, cannot be determined from the available evidence.

Pathogens Across the Group

The metagenomic analysis found traces of multiple pathogens beyond M. leprae. Hepatitis B virus was detected in individual s.167 with high confidence, and lower-coverage signals were validated in three others; genotyping recovered genotypes A and D, both known from medieval European HBV diversity. Human parvovirus B19, the cause of fifth disease, was recovered from four individuals as partial genomes clustering within genotype 2, with trace-level signals in five more. The parvovirus genomes help fill a temporal gap in the European B19V record between roughly 950 CE and the 20th century, and their assignment to genotype 2 extends the known persistence of that lineage.

In at least 3 of the 13 individuals analyzed, the team identified molecular evidence of infection by primary pathogens. Including lower-coverage HBV signals brings that number to 5, or approximately 38% — comparable to rates reported from broad-spectrum pathogen screening at other medieval sites. Including Streptococcus pneumoniae (which can be carried asymptomatically) and a periodontal pathobiont, Parvimonas micra, detected in individual s.133, puts the count at 7 individuals, or about 54%.

It’s a small community, and more than half of the people whose DNA could be recovered were carrying some detectable pathogen.

What the Island Was

Ibiza before 902 CE barely existed in the written record. By the time these 13 people were buried — between 950 and 1150 CE — the island had become a node in a network that extended south to the Sahel, east to the Levant, and north into a Mediterranean world increasingly structured by the confrontation between Islamic and Christian polities.

The genomes reflect this. They show an initial colonization wave from mainland al-Andalus, dominated by Imazighen with some Arab and Islamized Iberian participants, followed by a second pulse after the Almoravid conquest of Mallorca in 1115–1116 CE that brought new arrivals from North Africa, the Iberian peninsula, and — demonstrably — the far side of the Sahara. The timing estimates suggest admixture that had been ongoing for generations by the time these individuals were alive, some of it recent enough that the chromosomal blocks of North African ancestry were still long and largely intact.

The two individuals with the deepest European ancestry — genetically local, culturally Muslim — had apparently converted without marrying into the arriving populations. Or at least their ancestors had. And one of them, without any visible sign of disease, was carrying Mycobacterium leprae genotype 2F, a lineage traveling from the Mediterranean to Scandinavia and back again through the bodies of medieval Europeans who left few records other than this.

Further Reading

• Olalde, I. et al. “The genomic history of the Iberian Peninsula over the past 8000 years.” Science 363, 1237–1241 (2019).

• Oteo-Garcia, G. et al. “Medieval genomes from eastern Iberia illuminate the role of Morisco mass deportations in dismantling a long-standing genetic bridge with North Africa.” Genome Biology 26, 108 (2025).

• Pfrengle, S. et al. “Mycobacterium leprae diversity and population dynamics in medieval Europe from novel ancient genomes.” BMC Biology 19, 1–18 (2021).

• Schuenemann, V.J. et al. “Ancient genomes reveal a high diversity of Mycobacterium leprae in medieval Europe.” PLOS Pathogens 14, 1006997 (2018).

• Mühlemann, B. et al. “Ancient human parvovirus B19 in Eurasia reveals its long-term association with humans.” PNAS115, 7557–7562 (2018).

• Kocher, A. et al. “Ten millennia of hepatitis B virus evolution.” Science 374, 182–188 (2021).

• Fortes-Lima, C.A. et al. “Population history and admixture of the Fulani people from the Sahel.” American Journal of Human Genetics 112, 261–275 (2025).

• Dury, G. et al. “The Islamic cemetery at 33 Bartomeu Vicent Ramon, Ibiza: investigating diet and mobility through light stable isotopes in bone collagen and tooth enamel.” Archaeological and Anthropological Sciences 11, 3913–3930 (2019).

• Bennison, A.K. The Almoravid and Almohad Empires. Edinburgh University Press, 2016.

• Rodríguez-Varela, R. et al. “Five centuries of consanguinity, isolation, health, and conflict in Las Gobas: A Northern Medieval Iberian necropolis.” Science Advances 10 (2024).

The individual was male. He was probably about 110,000 years old. He had never been named or catalogued as a skeleton. He was, effectively, nobody — until he became one of the most informative people from the Pleistocene.

Two studies published this week in PNAS,1 2 both drawing on new Neandertal genomic data, arrive at a picture of Homo neanderthalensis that is striking for how fragile it makes them look. Not just at the end, in the final few thousand years before their disappearance, but throughout their entire existence across Eurasia. They were a species running on fumes.

The Fracture Lines

The Denisova Cave individual — referred to in the Massilani et al. paper as Neandertal D17 — belonged to a population of Eastern Neanderthals more closely related to another individual from the same cave, roughly 120,000 years old and female, than to any Neandertal from Western Europe. This isn’t surprising on its face: they came from the same place. What is surprising is the degree of genetic separation between these eastern Altai Neanderthals and the western populations in Europe.

The standard measure for comparing population differentiation is FST, which runs from 0 (identical) to 1 (completely separated). Among living humans, the most genetically distant populations on earth — Central African forest-dwelling groups like the Mbuti compared to Papuan Highlanders — reach an FST of around 0.27. These are populations that diverged somewhere between 130,000 and 220,000 years ago and have been largely separate ever since.

The gap between Eastern Neanderthals (D5 and D17 from Denisova Cave) and Western Neanderthals (represented by individuals from Vindija Cave in Croatia and a newly sequenced genome from Belgium) comes out at FST = 0.30. That is larger than the most differentiated living human populations. The Eastern and Western Neandertal lineages diverged only about 115,000 years ago, compared to 260,000 to 440,000 years of separate drift between Mbuti and Papuans. Neanderthals were differentiating faster. The Massilani team attributes this to smaller effective population sizes, which amplify genetic drift: when groups are small and isolated, allele frequencies shift quickly just by chance, without any particular selective pressure driving them.

How small were these groups? The genome of D17 shows that about 24 percent of his DNA sits in long homozygous runs — stretches where both copies of a chromosome are identical, a signature of recent inbreeding. In D5, the older female from the same cave, the figure is 20 percent. The Denisovan individual from the same site sits at 4 percent. Early modern humans come in between 1 and 6 percent. Population modeling using the length and distribution of these homozygous tracts suggests that the Altai Neanderthals — both the older Eastern ones and a later, Western-derived individual named Chag8 from a different cave in the same region — were living in groups of fewer than 50 individuals under realistic migration scenarios. The later Western Neanderthals in Europe appear to have lived in somewhat larger groups, but not dramatically so.

Princeton geneticist Joshua Akey, who was not involved in either study, put it plainly: the global Neandertal population was probably only a few thousand breeding individuals, spread across most of Eurasia.

There is something almost incomprehensible about that. Modern humans, for all the times we almost went extinct, have never been that reduced on a global scale while simultaneously occupying that much space.

The Bottleneck You Can See Coming

The second study, led by Charoula Fotiadou and Cosimo Posth at the University of Tübingen, approaches Neandertal demography from the other direction: not nuclear genomes from individuals, but mitochondrial DNA from dozens of specimens spanning the last 130,000 years of Neandertal history in Europe. Mitochondrial DNA is maternally inherited and present in large quantities in ancient bone, which makes it tractable even from fragmentary material. The team generated ten new mtDNA sequences from six sites in Belgium, France, Germany, and Serbia, and analyzed them against 49 previously published sequences.

The pattern they found is stark. Before about 75,000 years ago, European Neanderthals showed multiple distinct mtDNA lineages. Individuals from different sites across the continent sat on different branches of the phylogenetic tree. There was genetic diversity — not a lot by modern human standards, but genuine regional variation.

Then it collapses.

The vast majority of Late Neanderthals, those living after about 57,000 years ago, cluster within a single mtDNA lineage. From Iberia to the Caucasus, across the whole surviving range of European Neanderthals, almost everyone is descended in the maternal line from the same ancestral population. The analysis places the origin of that lineage at around 65,000 years ago, with a 95 percent confidence interval running from 56,000 to 76,000 years ago. And it pinpoints the likely geographic source: southwestern France.

What happened in between is not hard to guess. Marine Isotope Stage 4 — the glacial period peaking roughly 73,000 to 60,000 years ago — was cold and dry across Europe. The Fotiadou team’s analysis of the archaeological record, drawing on the ROCEEH Out of Africa Database (ROAD), shows Neandertal sites becoming dramatically concentrated in southwestern Europe during this period. The hotspot visible in the data through the 70,000 to 60,000 year window sits in southern France. Structured statistical tests on the spatial data show this cannot be explained by research bias or uneven sampling. Something pulled Neanderthals into that corner of the continent.

The genetic and archaeological evidence converge on the same scenario: a population that had been distributed across most of Eurasia contracted into a refugium. Most of the earlier diversity was lost. Whatever lineages existed outside that southwestern European core either died out or were so reduced as to leave no detectable genetic trace in later populations. The only exceptions are two individuals in France — one from Les Cottés, one from Grotte Mandrin — whose mtDNA sits outside the main Late Neandertal lineage, suggesting the refugium itself preserved at least some of the older variation.

When the ice retreated, sometime after 65,000 years ago, the survivors spread out again. Neanderthals reappeared at sites across Europe and into the Caucasus. But they were now, in the maternal genetic record, essentially one people. University of Tübingen paleogeneticist Cosimo Posth, a co-author on the Fotiadou study, described it directly: all the genetic diversity visible in the mitochondrial record before 60,000 years ago disappears, and a single lineage survives.

This is not extinction followed by replacement from somewhere outside. There is no external source population arriving from Asia or Africa. The Neanderthals who recolonized Europe after the glacial maximum were the descendants of the ones who had sheltered in southwestern France. What the bottleneck destroyed, you cannot get back.

42,000 Years Ago

The Bayesian skyline analysis in the Fotiadou study tracks effective population size through time, and the trajectory at the end is what you might expect given everything else in the data. There is no gradual thinning out. The line holds roughly steady until around 44,500 years ago, then drops sharply. It reaches its minimum around 42,000 years ago.

That timing overlaps with the arrival of anatomically modern humans in Europe, which most evidence places between 45,000 and 43,000 years ago, and with the sharp climatic instability of Greenland Interstadial transitions around the same period. Qiaomei Fu, a geneticist at the Chinese Academy of Sciences not involved in either paper, noted that for a population already as constrained as the Neanderthals had become, environmental volatility would have been especially dangerous. When you have no buffer — no large connected populations to draw migrants from, no reservoir of genetic diversity to draw on — each stochastic shock matters more.

Within roughly 3,000 years of that decline beginning, the Neandertal genetic signal disappears from the record entirely. Except, of course, for what survived through admixture with modern humans — the 1 to 4 percent of Neandertal ancestry still present in the genomes of everyone alive today outside sub-Saharan Africa.

Something else emerges from the Denisova Cave genome. The two oldest individuals from that site — D5 (around 120,000 years old) and D17 (around 110,000 years old) — both carry segments of DNA that trace to Denisovan ancestry. The locations of Denisovan-like segments in the two genomes are significantly correlated, suggesting they shared at least some of the same introgression events. But the later Neandertal from the region — Chag8, from nearby Chagyrskaya Cave, dating to around 80,000 years ago — shows no comparable Denisovan signal. Neither does the Western European Neandertal from Vindija. The Massilani team’s interpretation is tentative but suggestive: the Western-derived population that replaced the Eastern Neanderthals in the Altai sometime between 110,000 and 80,000 years ago may have been recent arrivals with no prior contact with Denisovans, or their ancestors may have arrived after whatever direct interactions the earlier populations had experienced.

It is a strange thing to contemplate: two groups of Neanderthals, separated by perhaps 10,000 years, occupying the same cave system at different times, genetically distinct enough from each other to be classified into separate regional populations, each leaving a different archaeological footprint, each with different admixture histories. And none of them aware that in some broader sense they were all running out of time.

Max Planck geneticist Hugo Zeberg, a co-author on the Massilani paper, framed the comparative value of these datasets in terms of natural experiments. Modern humans, Neanderthals, Denisovans: three lineages navigating the same world during overlapping time periods. The question of why one of them is still here is not answered by these papers, but it gets sharper. Neandertal populations, the data suggest, were almost never large enough to absorb disruption the way modern human populations apparently could. The differentiation exceeded anything seen in living humans today — and it happened faster, over shorter timescales, because the groups were smaller and more isolated. Drift overwhelmed everything else.

The bone fragment from Denisova Cave is 2.5 centimeters long. From it, researchers reconstructed a genome that shows a man living in a population of fewer than 50 people, already genetically separated from the people his kind would later become, already carrying DNA from another species entirely. He was part of a group at the far eastern edge of its range, in a mountain cave in Siberia, 110,000 years ago.

Whether that feels like isolation or just ordinary life is a question the data cannot answer.

Further Reading

• Hajdinjak, M., et al. (2018). Reconstructing the genetic history of late Neanderthals. Nature 555, 652–656.

• Mafessoni, F., et al. (2020). A high-coverage Neandertal genome from Chagyrskaya Cave. PNAS 117, 15132–15136.

• Prüfer, K., et al. (2014). The complete genome sequence of a Neanderthal from the Altai Mountains. Nature 505, 43–49.

• Prüfer, K., et al. (2017). A high-coverage Neandertal genome from Vindija Cave in Croatia. Science 358, 655–658.

• Slimak, L., et al. (2024). Long genetic and social isolation in Neanderthals before their extinction. Cell Genomics 4, 100593.

• Skov, L., et al. (2022). Genetic insights into the social organization of Neanderthals. Nature 610, 519–525.

• Higham, T., et al. (2014). The timing and spatiotemporal patterning of Neanderthal disappearance. Nature 512, 306–309.

• Urciuoli, A., et al. (2025). Semicircular canals shed light on bottleneck events in the evolution of the Neanderthal clade. Nature Communications 16, 972.