Anthropology.net

Two Ghosts in the Genome

Thu, 30 Jul 2026 20:56:37 GMT

Every so often a paper shows up that makes you recalculate what “human” has meant, genetically, for the last million years. This is one of them.

The basic problem researchers have wrestled with for over a decade is this: you can only find archaic DNA in living people if you have something to compare it to. Neanderthal ancestry was detectable because scientists sequenced actual Neanderthal genomes, pulled from actual Neanderthal bones. Same for Denisovans, though there we’ve only ever had one high-quality genome, from a finger bone in a Siberian cave. Without a reference genome sitting there for comparison, an unknown ancestor’s genetic contribution is essentially invisible. It just looks like ordinary human variation.

That’s a real limitation, because the fossil record and the genetic record have been quietly disagreeing with each other for years. Geneticists studying African populations kept noticing haplotypes, stretches of DNA inherited as a block, that didn’t fit any known demographic story. They looked old. Too old. But with no archaic African genome to check them against (the oldest African genome we have is under 20,000 years, and DNA doesn’t survive well outside permafrost), there was no way to confirm what these signals actually were.

Using genealogical relationships in DNA, TRACE uncovers hidden contributions from extinct human populations that left no sequenced genomes. Credit: Meaghan Marohn, https://meaghanmarohn.wordpress.com/

A team at UC Berkeley, led by Priya Moorjani along with graduate student Yulin Zhang and Johns Hopkins postdoc Arjun Biddanda, built a workaround.1 Instead of needing an ancient genome to compare against, their method, called TRACE, reconstructs the genealogical relationships buried inside modern genomes themselves. It reads what’s known as an ancestral recombination graph, essentially a map of how every stretch of DNA in a population traces back through shared ancestors over time. When a segment of DNA came from a deeply diverged, long-extinct lineage, it leaves a specific signature in that graph: branches that reach unusually far back in time, and haplotypes that stayed intact for longer than you’d expect from ordinary population history.

Applying TRACE to whole genomes from the 1000 Genomes Project, the team found two lineages nobody had a name for.

Two Bodies From Chile Just Rewrote the Genetic History of Smallpox in the Americas

Thu, 30 Jul 2026 20:46:16 GMT

Two skeletons from a burial site in northern Chile called Camarones 9 have been sitting in a museum collection for years, mummified naturally in the Atacama’s punishing aridity, catalogued as part of a Late Period Inca-affiliated community with signs of Spanish contact. Nobody was looking for a virus in them. The bone samples had been collected to study human DNA, not pathogens. But when researchers1 ran the femur fragments through a pathogen-screening pipeline, two of the thirteen individuals came back positive for variola virus, the cause of smallpox.

Credit: Unsplash/CC0 Public Domain

That’s the kind of find that sounds almost too clean, and for a while, epidemiological history has run on exactly this kind of narrative without the molecule to back it up. We’ve had the written record: Columbus arrives in 1492, a documented outbreak hits the Caribbean in 1518, the disease reaches Mexico by 1520, spreads down through the Isthmus of Panama into Inca territory by 1525, and eventually finds its way to Chile, with a well-documented 1561 outbreak generally marking its confirmed regional arrival. What we haven’t had, until this study, is the virus itself, isolated from a body that lived through it, with a genome that can be dated, sequenced, and placed on a family tree.

That’s what makes the Camarones 9 genomes matter. They’re not just old. They’re the first ancient smallpox genomes ever recovered from the Americas, full stop.

Why Ice Age Hunters Kept Hunting the Female Mammoths

Thu, 30 Jul 2026 20:41:54 GMT

Bones don’t come with a note explaining how they got there. That’s the problem that has dogged mammoth bone accumulations for decades: the huge piles of skeletal material found at sites like Kraków Spadzista in Poland or Dolní Věstonice in the Czech Republic, sometimes representing dozens or even over a hundred individuals, could be the leftovers of human hunting and butchery, or they could be a geological accident. A sinkhole. An alluvial deposit. A place where mammoths simply kept dying for reasons that had nothing to do with people.

Cut marks help. Embedded projectile points help more. But cut marks and stone points get you evidence at the level of individual bones, not entire assemblages, and taphonomists have spent a long time arguing about whether the marks they can see are frequent enough, and clear enough, to settle the question definitively for a whole site.

Excavation in Langmannersdorf 2025. Credit: Marc Händel/Austrian Academy of Sciences

A new approach1 sidesteps the argument almost entirely, and it does it by asking a different question: not “was this bone cut,” but “was this animal male or female.”

Provisioning Yengi-kent: How Sheep Teeth Reveal an Urban Survival Strategy on the Aral Sea’s Silk Road Frontier

Thu, 30 Jul 2026 16:24:07 GMT

Dzhankent sits on the northeastern shore of the Aral Sea, in a stretch of Kazakhstan that gets maybe five inches of rain a year and swings from -40°F winters to summers pushing 80°F. It is not a place that forgives sloppy food planning. And yet by the 10th century CE, this settlement, known to Arab geographers as Yengi-kent, had become the fortified capital of the Turkic Oghuz polity, a walled city dense enough with foreign and local residents that Khorezmian architects were brought in to redesign the whole interior.

Cities like that need feeding constantly, not seasonally. That’s the puzzle a team1 led by Kara Larson and Alicia Ventresca-Miller set out to solve, using a method that sounds almost too simple to work: cutting into sheep teeth.

Sheep mandibles were analyzed for isotopes in an international study including University of Michigan researchers. The study examined flexible herding techniques in a city in Kazakhstan during the 9th and 10th centuries. Credit: Alicia Ventresca-Miller, University of Michigan

Here’s the logic. As a sheep eats and drinks over its lifetime, the isotopes in its food and water get locked into the enamel of its molars, layer by layer, like tree rings. Carbon isotopes track what kind of plants it was grazing (different photosynthetic pathways leave different chemical signatures) while oxygen isotopes track the water it drank, which itself shifts with the seasons since summer precipitation and winter precipitation carry different oxygen signatures. Drill a series of samples along a single tooth, from the root to the chewing surface, and you get a chronological record. Not a broad average of what the animal ate over its whole life, but a month-by-month diary.

Location of Trench 2 on a Digital Elevation Model (DEM) map of Dzhankent (provided by H. Härke and I. Arzhanteseva). Credit: Archaeometry (2026). DOI: 10.1111/arcm.70199

The team applied this to the second molars of 16 sheep (Ovis aries) recovered from Trench 2 at Dzhankent, the deepest and most thoroughly dated excavation at the site, running through the 9th and 10th centuries. What they found was not one herding strategy but six.

Two Thumbnail-Sized Birds, Carved 40,000 Years Ago

Thu, 30 Jul 2026 16:01:29 GMT

Somebody sat in a cave in the Swabian Jura 40,000 years ago and carved a bird small enough to lose in a coat pocket. Not a rough approximation of a bird. A specific one, with a beak angled a particular way, with feather markings incised into a strip of mammoth ivory barely two centimeters long. Whoever made it was working at a scale where a single slip of the tool would have ruined hours of effort, and they were doing it during the last ice age, using stone blades, by firelight or daylight filtering into a cave mouth in what’s now southwestern Germany.

Two bird figurines carved from ivory found at the site of Hohle Fels Cave, approximately 40,000 years old. Credit: Ralf Ehmann / University of Tübingen, Grafik: urmu

Archaeologists from the University of Tübingen recovered two such figurines at Hohle Fels Cave in 2025, found1 about a meter and a half apart in the same archaeological layer. Both are carved from mammoth ivory. Both come from the Aurignacian, the technocultural period associated with the first modern humans to settle the Upper Danube region, and the layers at Hohle Fels have already produced some of the oldest figurative art known anywhere. The site’s reputation rests partly on the Venus of Hohle Fels and on flutes made from bird bone and ivory that count among the earliest musical instruments ever found. Now add two birds to the list, and these are unlike anything the site had produced before, not because of what they depict but because of how small they are.

At roughly 2 x 1 x 0.5 centimeters, they’re a fraction of the size typical for figurines from this period, which usually run three to nine centimeters. One weighs 1.3 grams. The other, incomplete, weighs 0.7. You could balance either on a fingertip and barely feel it. And yet the detail holds up under that constraint. The complete figurine, showing a bird settled low with its body pressed down, has clearly rendered eyes. The other, missing part of its body but preserving a full head and outstretched wings, has a distinctive beak and lines cut to suggest feathering. Nicholas Conard, who directs the excavations, put it plainly: this isn’t stylization. The carver recorded specific, deliberate features, the kind that let you ask what species of bird this is and what it’s doing, rather than just registering that it’s a bird at all.

A Civilization for Three Million People Left Almost No Trace Above Ground

Wed, 29 Jul 2026 21:02:19 GMT

Fly a laser over enough rainforest and the forest starts to admit things. Not all at once. The canopy has to be pierced pulse by pulse, from a plane running long straight lines a few hundred meters up, and even then the ground only shows itself in fragments, wherever enough light manages to slip through gaps in the leaves and bounce back. It is a slow, expensive, unglamorous way to look at the earth. It is also, as of this year, how a team led by Martti Pärssinen has forced1 a rewrite of what people are willing to say about the precolonial Amazon.

The rewrite is this: between roughly 600 BCE and 850 CE, a network of interconnected societies in what is now the Brazilian state of Acre and neighboring Amazonas built somewhere between 24,000 and 30,000 earthworks across an area of about 183,000 square kilometers, roughly the size of Syria. At its peak, around 100 to 300 CE, this civilization may have supported somewhere between 1.25 and 3 million people. Pärssinen and his colleagues call them the Aquiry, after the Indigenous name for the Acre River, and their paper, published in Nature, reads like a demolition of the polite skepticism that has followed Amazonian archaeology for decades.

Fazenda Atlântica archaeological site in Acre, Brazil. Credit: M. Pärssinen & F. de Novaes

The number that ought to stop you is the multiplier. When the researchers flew LiDAR over ground they had already surveyed using satellite imagery, they found five times more earthworks than the satellite data had shown. Where visible-light imagery from a deforested patch had picked out 30 structures, the laser found 156, at the exact same site. Extend that correction across the whole known extent of Aquiry territory and you get a civilization roughly ten times larger than anyone had previously proposed.

None of this comes from grand stone monuments, because there is no stone. Southwestern Amazonia sits on a landscape without the raw material that let other precolonial societies build cities that read as cities to modern eyes. What the Aquiry left instead are earthworks: ditches cut nine to thirteen meters wide and up to five meters deep, banked with the excavated soil, laid out in squares, rectangles, and circles that sometimes cluster into complexes of six or eight enclosures joined by wide roads. The largest known example covers fifty hectares. Most are smaller, under a hectare and a half on average, but there are thousands of them, and until this survey most sat buried under forest that satellite cameras simply could not see through.

What Flat Heads Tell Us About a Million Years of Difficult Babies

Wed, 29 Jul 2026 14:33:56 GMT

Every pediatrician has a name for it now: flat head syndrome, or deformational plagiocephaly, if you want the clinical term. It’s the mild, usually harmless flattening that shows up on the back or side of a baby’s skull after months of lying on a soft flat surface, unable to turn its own head against gravity. Parents worry about it, then mostly forget about it once their kid starts walking around with a full head of hair. It seems like exactly the kind of thing that has nothing to do with deep time.

It turns out to be exactly the kind of thing that has everything to do with deep time.

A team led by Yousuke Kaifu at the University of Tokyo has just published a study in Proceedings of the Royal Society B1 arguing that skull flattening is not just a modern parenting footnote. It is a physical signature of a trait so central to being human that its absence or presence in a fossil skull can tell you something about how that individual’s infancy went, and by extension, whether someone had to take care of it.

Back views of the skulls of a historical human (left) and a Homo erectus from Ngawi, Java (right: undated), exhibiting severe transverse deformational plagiocephaly. Credit: Kaifu Y et al. 2026 CC BY

Start with the strange fact this whole study is built on. Human babies are almost comically underdeveloped compared to other great apes. A newborn chimpanzee can lift its head within weeks and cling unaided to its mother’s fur by two months old, freeing her to move through the forest without using her hands to hold the infant in place. A newborn human can do neither. Our necks are weak, our skull bones are thin and mobile, held together by soft fontanelles that won’t fully close for well over a year, and our brains are still doing roughly three quarters of their total postnatal growing after we’re already out in the world. Chimp and macaque brains, by contrast, do most of their growing before birth, arriving already past the steepest part of the growth curve.

The crosses on these graphs represent the great apes (violet = P. troglodytes, blue = P. paniscus, orange = Pongo, black = Gorilla), while the light blue circles = H. sapiens, and red stars = fossil Homo (LB = Liang Bua, Ng = Ngandong, Sm = Sambungmacan). The further the point is from 0 on the vertical axis, the greater the distortion. On the right, the density curves show the distributions of the pooled great ape (dotted line) and modern human (solid line). Credit: Kaifu Y et al. 2026 CC BY

The going explanation for this has long been the obstetrical dilemma: bipedalism narrowed the pelvis, big brains demanded a wide birth canal, and something had to give. Babies got evicted early, before their skulls and nervous systems caught up, because keeping them in utero any longer risked a birth canal that couldn’t accommodate a bigger head. It’s a tidy story. The problem has always been figuring out when in the hominin lineage this trade-off actually kicked in, since infant fossils are vanishingly rare and the skeletal evidence that does survive is fragmentary.

The Dead Who Lived With You

Wed, 29 Jul 2026 14:27:55 GMT

For most of the twentieth century, if an anthropologist wanted to understand how ancient people related to their dead, the move was simple: go find a living society that still does something similar, and use it as a model. This is called ethnographic analogy, and it has quietly shaped how archaeologists interpret bones under floors, grave goods in courtyards, and the layout of settlements themselves.

Sabina Cveček ran into a problem with this approach almost by accident. She was building a database of residential burial, the practice of burying people inside or immediately next to the homes they lived in, using ethnographic records collected by anthropologists over the past two centuries. Working through 60 societies drawn from the eHRAF World Cultures database, she found that only 21% of them practiced residential burial. That number felt wrong. Archaeologists kept finding graves under house floors at site after site. Why weren’t living societies doing the same thing?

Dr. Sabina Cveček visiting the site of Lepenski Vir in the Danube Gorge in Serbia. Credit: Sabina Cveček

So Cveček, along with Carol Ember at Yale’s Human Relations Area Files, Timothy Earle at Northwestern, and Gary Feinman and William Parkinson at the Field Museum, built a second dataset. This one pulled from 49 archaeological traditions spanning the globe, everything from Neolithic Anatolia to the Classic Maya to Bronze Age Iran. The result:1 43% of those archaeological cases showed residential burial. Double the ethnographic rate, and the difference held up statistically.

The Lepenski Vir (6300–6000 BC) architectural phase, characterized by trapezoidal buildings in which numerous infant burials were placed close to the house walls. Credit: Sabina Cveček

That gap is the whole story here, and it’s a bigger deal than it might first sound.

Ancient Genomes from a Cave Above the Clouds Reveal an Unknown Himalayan Population

Wed, 29 Jul 2026 04:12:57 GMT

There is a cave in Ladakh called Abi Srinjamo Bao, the Old Lady Spider Cave, sitting in pastureland above the village of Yogma Kharbu at an altitude where the air itself starts to feel thin and unfamiliar. Inside, past a narrow tunnel, bones are scattered across the floor of a second chamber in two dense clusters, some of them still carrying scraps of mummified skin, ligament, and muscle. People have been walking into this cave for centuries, which means the skeletal material inside it has been disturbed, shuffled, picked through. But the bones themselves turned out to hold something none of that disturbance could erase: a genetic signature that doesn’t match anything common in South Asia today.

The Old Lady Spider Cave. (A) Plan and section of the cave, drawn from a previously unpublished point cloud shared with permission by D. Rosati. Orange shading shows where human remains were found, while gray shading shows a part of the cave covered by a roof collapse. (B) Previously unpublished photograph of bones in room 2, taken by the authors during the excavation. Credit: Science Advances (2026). DOI: 10.1126/sciadv.aeb3636

A team led by researchers from Harvard and the Deccan College Post-Graduate and Research Institute extracted DNA from ten individuals recovered during a 2021 excavation, radiocarbon dated to roughly 1,500 years ago. What they found1 was startling in its simplicity. Each of these people carried ancestry that split almost exactly down the middle: about half tracing to a population resembling present-day North Indians, the other half to a population closely related to ancient Tibetans. That kind of near-even split is rare among South Asians living today. A handful of modern Himalayan groups, the Minero and Balti of Ladakh among them, come close. Almost nobody else does.

What 148 Skeletons from an Iron Age Cemetery Reveal About Who Really Ate Well

Wed, 29 Jul 2026 04:09:14 GMT

Novilara sits inland from the Adriatic coast in what’s now the Marche region of Italy, and between the 8th and 7th centuries BCE it was one of the largest cemeteries of the Picene culture, a group whose material world has always been easier to read than its people. Excavators have been pulling burials out of Novilara since the 19th century, more than 400 in total across two campaigns a century apart, and the graves themselves are eloquent. Amber earrings. Bronze brooches. Spindle whorls and weaving tools tucked beside the bodies of women. Weapons beside men. A carved stone stele showing ships and armed figures, mysterious enough that archaeologists still argue over what it commemorates. Some burials cluster together, richly furnished, as if kin groups were staking a claim to status that would outlast them.

None of that tells you what anyone actually ate.

Zita Laffranchi. Credit: Zita Laffranchi

A new study in the Journal of Archaeological Science,1 led by Zita Laffranchi and Marco Milella with colleagues from the universities of Córdoba, Pisa, Bern, and Sapienza, went looking for that missing layer. The team sampled bone collagen from 148 individuals at Novilara and measured three stable isotopes: carbon, nitrogen, and sulphur. Carbon separates plants that photosynthesize differently, mainly wheat and barley versus millet-type crops. Nitrogen climbs with each step up the food chain, so it tracks how much animal protein someone was eating. Sulphur does something else entirely: it flags where someone’s food, and by extension where they, came from, because sulphur signatures shift with proximity to the sea and with local geology.

Put together, these three numbers turn a skeleton into a kind of diary, written not in words but in diet, one that nobody at the time could have edited for effect.

Malta’s Impossible Mesolithic

Wed, 29 Jul 2026 04:06:49 GMT

Here’s the problem with Malta. Somewhere between 8,500 and 7,500 years ago, hunter-gatherers were living there. We know this now from excavations at Latnija, in the island’s north, where stone tools and butchered animal bone push the human presence back two thousand years earlier than anyone expected. That much has already reshaped the story of Mediterranean prehistory. But a new study in PNAS Nexus,1 led by James Blinkhorn and colleagues, asks a different question, one that’s arguably more unsettling: even granting that foragers reached Malta, how did they stay there?

Cave site of Latnija in the northern Mellieħa region of Malta, which revealed evidence of over 1000 years of forager occup

The island is small. It’s dry. It has almost no large terrestrial fauna beyond an endemic deer population, and it sits across open water from Sicily, its nearest neighbor. None of this is new information. What’s new is a quantitative attempt to figure out exactly how many people Malta’s environment could have supported at any given moment between the Last Glacial Maximum and the present, and whether that number was ever large enough for a population to survive on its own, without outside contact, for the better part of a millennium.

The answer is no. Not even close.

Every Excavation Is a Deadline You Didn’t Know You Had

Tue, 28 Jul 2026 16:15:35 GMT

Dig a trench and you destroy it. That’s the part nobody outside archaeology quite absorbs. You can’t re-excavate a stratigraphic layer to check your notes. Once the soil comes out, the context is gone, and whatever you wrote down, photographed, or scanned on the way down is now the only version of that moment that will ever exist. This isn’t a metaphor about the fragility of the past. It’s a description of a Tuesday on site.

A new study1 out of the University of Barcelona sat down with fourteen principal investigators running archaeological projects across Catalonia and asked them, in long, unhurried interviews, how they actually manage all that irreplaceable material. Not how they’re supposed to, according to some funding agency’s checklist. How they actually do it, in practice, on the ground, with the tools and time they have.

Credit: Unsplash/CC0 Public Domain

The answer turns out to be: pretty well, mostly by instinct, and almost entirely without anyone having taught them how.

When Your Ear Betrays Your Mouth: What Mandarin Speakers’ English Vowels Reveal About the Wiring Between Hearing and Speech

Tue, 28 Jul 2026 16:10:32 GMT

Ask a Mandarin speaker to say “bit” and “beat” in quick succession, and something curious happens. The vowels come out different, clearly different, but not necessarily in the way a native English speaker would make them different. English distinguishes these two vowels mostly through vowel quality: where the tongue sits in the mouth changes the sound’s spectral shape. Length matters too, but it’s the minor cue. Mandarin speakers often flip that hierarchy. For them, duration becomes the main signal and quality fades into the background. The vowels are still contrasted. They’re just contrasted using the wrong tool, at least by native standards.

This kind of mismatch has been documented for decades. What hasn’t been settled is a more basic question: when a learner’s ear and mouth disagree about which acoustic cue matters, are the two systems actually talking to each other at all? If someone leans on duration when listening to a vowel contrast, does that same lean show up when they try to reproduce it? Or do perception and production operate as separate systems that happen to produce roughly compatible outputs?

Figure 4. Vowel imitations across three spectral steps in the F1–F2 space (Bark scale), shown as kernel density contours for /i, ɪ/ (left) and /æ, ɛ/ (right). Notes: Ellipses show participant imitations (Orange = Step 1; Gray = Step 4 ambiguous; Purple = Step 7). Matching triangles show the original target vowels.

A new study1 out of National Yang Ming Chiao Tung University, led by Will Chih-Chao Chang and Yu-An Lu, gets at this question with more precision than most previous attempts, and the answer turns out to be genuinely interesting: it depends on which sound, and which acoustic dimension you’re asking about.

Eight Sets of Footprints on a Vanished Shoreline

Tue, 28 Jul 2026 02:06:09 GMT

In 1978, Kay Behrensmeyer was watching Kenyan colleagues cut a trench through sediment layers near Lake Turkana when something round and deep caught her eye. It turned out to be a hippo footprint. When the crew cleared the surface around it, two more prints appeared, this time from something walking on two legs.

Behrensmeyer had no way of knowing it then, but she’d found the edge of something much bigger. The site got buried in sand for protection, the way these things often go, and it would be decades before anyone came back to see how far it extended. When they finally did, in 2016 and again in 2023, the answer1 turned out to be: much further than anyone guessed. The layer held 21 footprints from eight different individuals, laid down alongside tracks from hippos, antelope, and other animals moving across the same stretch of mud roughly 1.43 million years ago.

Overview of the excavated track surface with a mix of deeper hippopotamus and antelope footprints and shallower hominin tracks visible. Credit: Kay Behrensmeyer, Smithsonian.

The site is called GaJi10, and the group that made those prints has turned out to be something other than what the researchers expected.

Study lead Kevin Hatala, an associate professor of biology at Chatham University in Pittsburgh and an associate researcher at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, pointing out footprints of hominins among deeper hippopotamus tracks. Credit: Kay Behrensmeyer, Smithsonian.

What the Bones of Ordinary People Say About the Maya Diet

Mon, 27 Jul 2026 16:00:59 GMT

For a long time, if you wanted to know what the ancient Maya ate, you had to ask the rich ones. Skeletons from Tikal, Calakmul, Caracol, palace burials with jade and painted vessels. That bias made sense practically. Elite tombs get excavated first, get funded, get published. But it left a strange gap in the record: we knew a lot about how kings and nobles ate and almost nothing about everyone else, despite commoners making up the overwhelming majority of any Classic Maya population.

A new study in PLOS One1 fills part of that gap by going somewhere unglamorous. Southeastern Petén, in northern Guatemala, was never home to a single dominant city. It was a patchwork of smaller, loosely bounded polities, more than 400 documented sites scattered across mountains, river basins, and savanna, none of them centralized the way Tikal or Calakmul were. A team led by Gloria Ivonne Hernández Bolio sampled bone and teeth from 57 individuals across these sites, almost entirely from non-elite burials, and ran them through stable isotope analysis to reconstruct what these people were actually putting in their bodies

The archaeological site of Curucuitz, in southeastern Peten, Guatemala, where maize is still grown today. Lower image: The A-630 vessel from Ixtutz, depicting four individuals, arranged in pairs, sharing the contents of a pitcher. Credit: Proyecto Atlas Arqueológico de Guatemala, CC-BY

The short answer: maize, a lot of it, for centuries, without much variation by wealth or sex, and without the dietary shifts that show up elsewhere in the Maya world when things start to fall apart.

A Cancer Gene That Is Still Fighting Its Own Past

Fri, 24 Jul 2026 21:19:59 GMT

Somewhere on chromosome 17, in the DNA of tens of thousands of Han Taiwanese people, there is a stretch of BRCA1carrying sixteen variants that should not still exist. Fifteen of them are classified as pathogenic. Fourteen cause frameshifts, the kind of mutation that scrambles a protein’s reading frame past the point of repair. This haplotype is rare, present in roughly six of every ten thousand copies of the gene, and it is getting rarer. Purifying selection is doing what purifying selection does: quietly deleting a bad genetic sentence, one generation at a time.

But a few hundred kilobases away, in the same gene, in the same population, a completely different haplotype is thriving. It carries variants at a frequency of about 37 percent, common enough to be unremarkable, and it shows the unmistakable genomic signature of positive selection, the kind normally associated with something being actively favored rather than tolerated. Whatever is happening at BRCA1 right now, it is not one story. It is two, running in opposite directions, in the same gene, in living people.

That is the strange, quietly startling core of a new study in The American Journal of Human Genetics,1 led by Jing-Lian Chen and Wen-Ya Ko’s team at National Yang Ming Chiao Tung University, working with Yoko Satta at SOKENDAI. The tool they used to find this wasn’t a fossil, an ancient genome, or a cross-population comparison. It was a biobank, and a fairly simple idea about age.

Graphical abstract. Credit: The American Journal of Human Genetics (2026). DOI: 10.1016/j.ajhg.2026.07.002

The Baseline That Keeps Moving

Fri, 24 Jul 2026 21:01:03 GMT

About 80,000 years ago, a brown bear died in what is now Furninha Cave on the Portuguese coast. It weighed somewhere close to 385 kilograms, heavier than almost any brown bear alive in Europe today, built with a robustness that made researchers examining its bones think, more than once, of cave bears. Except there were no cave bears in Portugal. Decades of excavation have never turned one up. The cavities the country offers apparently weren’t large enough to support the multigenerational hibernation colonies the species needed, so Ursus spelaeus simply never moved in.

D. Estraviz-López with an ancient brown bear skeleton. Credit: D. Estraviz-López

That absence left something open. And a new study1 of over 500 bear bones from six Portuguese sites, led by Darío Estraviz-López at NOVA School of Science and Technology, suggests Iberian brown bears may have expanded into it, growing large and robust in a way that converges, at least superficially, on the cave bear body plan. It’s a hypothesis, not a settled fact, and testing it will need ancient DNA and isotope work still to come. But it points at something worth sitting with: the “natural” community of Pleistocene Portugal wasn’t a fixed cast of species filling predetermined roles. It was contingent, shaped by which species happened to be present, which happened to be absent, and how the ones present responded to the gap.

That contingency, and how badly it gets scrambled by human activity, is also the subject of a much larger review published the same year in Science Advances,2 led by Thomas Matthews at the University of Birmingham along with dozens of co-authors. Their focus is islands rather than caves, extinction and introduction rather than body size, but the underlying argument rhymes with the Portuguese bear story closely enough that reading them together changes how each one lands.

The World Once Had Tens of Thousands of Languages. Almost All of Them Are Gone.

Fri, 24 Jul 2026 20:33:49 GMT

In the highlands of Papua New Guinea, roughly 840 languages are spoken by a population of about 12 million people. Europe, with 745 million people, has a quarter of that number. Nobody who studies language finds this comparison surprising anymore. New Guinea’s fragmented terrain, its isolated valleys and self-sufficient communities, is the textbook case of what linguists call a hotspot: a place where small groups stay small, stay put, and drift apart linguistically because they don’t need each other.

What a team led by Damián Blasi, along with Marcus Hamilton, Russell Gray, and Claire Bowern, has now argued in Science1 is that New Guinea isn’t an anomaly. It’s a relic. A surviving fragment of a linguistic condition that used to be the global default, before most of the planet’s languages were erased.

1Contemporary languages and endangerment status. Made with a Natural Earth base map using data from Glottolog 5.3 and information adapted from GlottoScope. Credit: Claire Bowern

The Baboon That Glows Like the Moon

Thu, 23 Jul 2026 16:15:29 GMT

Harnufi the sacred baboon arrived in Egypt during the sixth year of Ptolemy V’s reign, was raised in a temple, and was buried thirty one years later, on September 6th of 168 BC. His entombment is recorded in the Baboon Catacomb at North Saqqara, one obituary among hundreds. The scale of that necropolis alone forces a strange question. Ancient Egypt had no wild baboons. None. The species is absent from the Holocene fossil record entirely. Every single mummified baboon interred in Egyptian soil, and there are hundreds of them, had to be imported, kept, fed, and eventually embalmed by people who had no local population to draw from and no ecological reason to care about the animal at all.

And yet baboons became one of the two bodies of Thoth, the god of writing, wisdom, and the moon. His other manifestation, the sacred ibis, at least makes intuitive sense. The bird was native, common, and its curved beak has an obvious resemblance to a crescent moon. The baboon offers no such easy explanation. Egyptologist Emily Teeter has called the whole arrangement inscrutable. A new paper by Nathaniel Dominy and Catherine Hobaiter, published in Time and Mind,1 tries to solve it with a spectrometer.

Thoth as baboon. Credit: Catherine Hobaiter

A Kangaroo Puppet, a Noongar Word, and the Grammar of Belonging

Thu, 23 Jul 2026 16:09:43 GMT

A four-year-old sees a photo on his father’s phone. It’s his dad, holding a goanna he caught himself. The boy points and says one word: kaarda. Not “goanna.” Not the English gloss the teacher had offered minutes earlier. He reaches for the Noongar word because that’s the word that fits the memory, the hunt, the pride in his father’s hands.

That single moment, recorded during a year-long study inside an Aboriginal playgroup in Boorloo (Perth), is doing a lot of quiet work. It’s evidence that a four-year-old already carries a working theory of when to use which language, and it’s the seed of an argument that researchers Sender Dovchin and Tamara Tucker are making about how language education gets structured in Australia, and by extension, what “language education” is even allowed to mean.

Associate Professor Sender Dovchin (left) and Noongar language educator Tamara Tucker have developed a new teaching model. Credit: Curtin University

Their study, conducted weekly for a year with preschool-aged children learning Noongar, the language of the traditional custodians of the Perth region, produced a new pedagogical framework they call translingual yarning pedagogy, or TYP. The name is clunky. The idea underneath it is not.