What they extracted, after nearly four years of grinding, boiling, dissolving and recrystallising, was about one-tenth of a gram of radium chloride. A speck. Less than the weight of a grain of rice.

The shed on the rue Lhomond

The workspace was not a laboratory in any modern sense. It was an abandoned dissecting room behind the Municipal School of Industrial Physics and Chemistry, with a glass roof that leaked when it rained and no fume hoods of any kind. In summer it baked. In winter the Curies wrote about their fingers going numb around the iron stirring rods.

Pitchblende is a dense, tar-black uranium ore. The Curies had reasoned, from careful electrometer measurements Pierre had refined using his own piezoelectric instruments, that pitchblende was more radioactive than the uranium it contained could account for. Something else had to be in there. Something rarer, and stronger.

Austria’s imperial government agreed to ship them the waste ore left over after uranium had been extracted for glassmaking. Eight thousand kilograms of it arrived on carts, dumped in the courtyard mixed with pine needles from the Bohemian forests. Marie processed it in twenty-kilogram batches.

“I had to spend a whole day mixing a boiling mass with a heavy iron rod nearly as large as myself,” she wrote later. “I would be broken with fatigue at the day’s end.”

What the work actually looked like

The chemistry, in outline, was brutal repetition. Dissolve pitchblende in hydrochloric acid. Precipitate the sulfides. Separate what remains active from what doesn’t. Then do it again with the active fraction. And again. And again.

Radium behaves chemically almost exactly like barium, which is why the ore contained radium at all and why isolating it was so agonising. Marie had to perform thousands of fractional recrystallisations, exploiting the tiny difference in solubility between radium chloride and barium chloride in hydrochloric acid. Each cycle concentrated the radium a little more. Each cycle took hours.

By 1902, she had her decigram. She measured radium’s atomic weight at 225, close to the modern value of 226. The substance glowed in the dark, warm to the touch, kicking off heat with no apparent fuel. Pierre carried a small vial of it in his waistcoat pocket to show visitors.

His fingertips, by then, were raw and inflamed. So were hers.

The dose they didn’t know they were taking

Nobody in 1900 understood ionising radiation the way a modern radiation-safety officer does. The Curies knew radium caused burns — Pierre deliberately strapped a sample to his arm to observe the lesion — and they knew it could destroy tumour cells, which is why radium therapy became one of the first cancer treatments. What they did not appreciate was cumulative exposure. Alpha particles inhaled as radon gas. Beta particles from decay products deposited in bone. Gamma rays passing through everything.

Marie stored radium samples in her desk drawer at home. She and Pierre described the shed years as “the best and happiest of our life,” and would return in the evenings to watch the tubes glow on the shelves like “faint fairy lights.”

She would die in 1934 of aplastic anemia, aged 66. Her daughter Irène, who worked alongside her at the Radium Institute, would die at 58 of leukaemia. Both illnesses are consistent with prolonged radiation exposure.

Why the notebooks are still hot

Radium-226 has a half-life of about 1,600 years. Which means the radium contamination Marie tracked on her fingertips into her lab notebooks in 1902 has, in the intervening years, decayed by less than five per cent. It is essentially as radioactive as it was the day she wrote in it.

The Bibliothèque nationale de France holds her papers in lead-lined cases. Researchers who want to consult them must sign a liability release and handle the pages with protective equipment. The same goes for her personal effects at the Musée Curie in Paris — furniture, chairs, cookbooks. A 2025 BBC feature followed radiation surveyors retracing the Curies’ movements around Paris, still finding contamination in the plaster and floorboards of buildings the couple had worked in more than a century ago.

Her body itself is buried in a lead-lined coffin. When her remains were transferred to the Panthéon in 1995 — the first woman interred there on her own merit — the coffin was shielded with lead lining because her bones remain measurably radioactive.

The scale of what one-tenth of a gram meant

To picture the ratio: eight tonnes of ore in, one-tenth of a gram out. That’s a concentration factor of roughly eighty million to one. If you started with a fully loaded articulated lorry of pitchblende, you would end with less material than fits on the tip of a pencil.

And yet that speck was enough to change physics. Radium was the first element whose radioactivity was so intense it could be studied directly. It gave Ernest Rutherford the samples he needed to work out alpha, beta and gamma radiation. It gave Frederick Soddy the evidence for isotopes. It gave medicine its first tool for treating deep-seated tumours. Every subsequent development in nuclear physics traces back, in some way, to that decigram in the Paris shed.

The Curies could have patented the extraction process. They refused. Radium, Marie said, belonged to science.

Two Nobels and a horse-drawn wagon

In 1903, the Nobel Committee awarded the physics prize jointly to Henri Becquerel and to Pierre and Marie Curie for their work on radioactivity — a word Marie had coined. She was the first woman ever to receive a Nobel. Neither Curie attended the ceremony; Pierre was ill and Marie was recovering from a miscarriage.

Three years later, on 19 April 1906, Pierre stepped off a kerb on the rue Dauphine in the rain and was struck by a horse-drawn dray. The wheel crushed his skull. He was 46.

The Sorbonne handed Marie his teaching post, making her the first woman ever to hold a professorship at the ancient university. In 1911 she won a second Nobel, this time in chemistry, for the isolation of radium and the discovery of polonium — named after her occupied homeland. She remains the only person to have won Nobels in two different sciences.

The women who came to the shed’s successor

After Pierre’s death, something else began to happen in Marie’s laboratory. Women scientists — barred from universities and posts across most of Europe and North America — started arriving in Paris to work with her. The Norwegian radiochemist Ellen Gleditsch. The Canadian nuclear physicist Harriet Brooks, who had already helped Rutherford identify what would later be called radon. Dozens of women in total passed through what came to be called simply “the Curie lab.”

The writer Dava Sobel described the moment she realised how many forgotten female chemists had trained there. “She had a room full of women,” Sobel said, “and nobody knows.”

Marie herself had been shut out of the University of Warsaw as a young woman because it did not admit women, and had studied instead at the underground “Flying University” before eventually reaching the Sorbonne. She understood exactly what her lab represented to the women who found their way to it.

What the leaky shed left behind

The shed on the rue Lhomond is gone. The Radium Institute she founded still stands, now part of the Institut Curie, one of the world’s leading cancer research centres. The gram of radium she brought back from the United States in 1921, after a fundraising tour arranged by the American journalist Marie Meloney, is still in the institute’s collection, stored under heavy shielding.

Radium itself has largely fallen out of use. Cobalt-60 and caesium-137 replaced it in medical therapy. The luminous radium paints once used on watch dials — the ones that killed the “Radium Girls” who licked their brushes to a fine point — were banned. Radium’s afterlife is mostly cautionary now.

But the notebooks endure. If you visit the Bibliothèque nationale and ask to see Marie Curie’s manuscript pages, a librarian will bring you a Geiger counter reading along with the request form. The needle jumps. The paper smells faintly of old ink and dust. The handwriting is small and careful, columns of numbers, sketches of glassware, marginal notes in a mix of French and Polish.

In 3,626 CE, when the radium she smeared into those pages has decayed by half, the notebooks will still tick. Whoever opens them then — if anyone does — will be reading a document that is, in a very literal sense, still emitting the discovery it records.

Between July and September 2017, a team from California Polytechnic State University buried speakers along roughly a third of a mile stretches of trail in the Boulder Open Space and Mountain Parks system, wired them to play recordings of birdsong, and waited to see whether hikers who walked past could tell the difference. Most, by the researchers’ own account, could not.

The study, published in Proceedings of the Royal Society B in December 2020, was led by Danielle Ferraro, then a biology graduate student at Cal Poly, with Clinton Francis, Zachariah D. Miller, Lauren Ferguson, Derrick Taff, Jesse Barber and Peter Newman as co-authors. The paper’s title calls the setup a “phantom birdsong chorus,” and that is the more precise term. It was not run at dawn. Data collection took place across the middle hours of the day, when hikers were actually out on the trails, and the recordings were built to simulate a denser, more varied mix of bird species than the trail would ordinarily produce at that time, not to recreate the specific acoustic conditions of first light. The “livelier dawn chorus” framing captures the spirit of the manipulation, the layering of many species’ songs into a chorus richer than what was really there, but the timing detail is worth keeping straight.

What the speakers actually did

The researchers alternated the trail sections between “on” blocks, when the hidden speakers played a mix of recorded songs from birds that could plausibly be heard in that habitat, and “off” blocks, when hikers heard only the ambient soundscape. The blocks rotated on a weekly basis across the two trail sections. Hikers were intercepted as they came off the trail and asked to complete a questionnaire about their experience, without being told in advance what the study was investigating.

Clinton Francis, who oversaw the project, told National Geographic that the deception was deliberate and that his team worried about it being discovered: “One of my graduate students was so nervous that an incredibly good eBirder was going to find a speaker and be outraged.” To the researchers’ knowledge, that never happened. The speakers went undetected for the length of the study, which is itself part of the finding: whatever effect the phantom chorus had did not depend on hikers consciously registering that something artificial was being piped into the landscape.

What hikers actually reported

The exposure window was short. Francis put it at seven to ten minutes; a later account of the study describes it as lasting “about 10 minutes.” In that window, hikers who walked through an “on” block reported higher restorative effects on their questionnaires than hikers who walked through the same stretch during an “off” block, on both trail sections. The effect was consistent across the two trails, but it worked through different routes. On one trail, hikers who heard more birdsong simply reported feeling better, without indicating that they thought more bird species were present. On the other trail, the improved sense of well-being was mediated by perceived biodiversity: hikers who heard the phantom chorus were more likely to believe more species lived along that stretch, and it was that belief, not the sound alone, that the researchers’ statistical model linked to the reported improvement.

This is one study, not settled consensus about birdsong and well-being generally. It is a controlled field experiment run on two trail sections in one Colorado park system over a single summer, using self-reported questionnaire data collected immediately after a short hike. The authors themselves frame it as adding to “a growing body of evidence linking mental health to nature experiences,” not as closing the question of why nature helps.

Why the perception mattered as much as the sound

The split between the two trails is the part of the paper that repays a second look. If the phantom chorus worked purely as a pleasant background noise, its effect should have looked the same on both trails: more birdsong, better mood, independent of whatever the hiker believed about the ecosystem around them. Instead, on one of the two trails, the effect ran through a hiker’s belief about biodiversity, meaning the sound worked partly by changing what people thought was true about the place they were standing in, not only by changing what they heard.

That distinction matters for how the result should be used. It is not evidence that ambient bird noise, piped in indiscriminately, produces a reliable calming effect on its own. It suggests, on this dataset, that the perception of a richer natural environment, formed on the basis of what a person can hear, is doing some of the work. Francis described the effect in blunter terms to National Geographic, calling it a “clearing of the mind,” and said he was “flabbergasted” that so short an exposure produced a measurable shift at all.

What the study does not show

The paper does not measure long-term outcomes. It says nothing about whether the restorative effect persists after a hiker leaves the trail, or whether repeated exposure compounds or fades. It relies on self-report, collected once, right after the walk, which is a real limitation of the design rather than a footnote: people are not always accurate narrators of their own mental states, particularly when asked to summarise them in a brief questionnaire under mild time pressure. And the study was carried out in a single protected area with a specific existing soundscape, hikers who chose to be there, and a demographic that is unlikely to represent hikers, or non-hikers, more broadly.

The authors’ own conclusion is narrower than the headline version of the finding. They argue that audition, the sense of hearing, is an underexamined route by which people register the biodiversity of natural environments and draw benefit from that environment, and that protected-area managers might treat noise pollution as something worth managing, not merely tolerating, if they want visitors to have a richer experience. That is a modest, specific claim about soundscapes and park management. It is not a claim that recorded birdsong is a reliable treatment for anxiety, or that any hiker exposed to more birdsong will feel calmer in a predictable, reproducible way.

The part that is easy to miss

What stands out about the design, more than the headline result, is what it required the researchers to give up. To test whether birdsong causes a change in how people feel, rather than merely correlating with it, they needed a controlled manipulation, which meant deceiving the people they were studying about what was actually happening around them. The alternative, telling hikers in advance that some of what they were hearing was recorded, would have compromised the very thing being measured. The paper’s contribution is less about birds than about method: proving a causal link between a single sense and a self-reported outcome in an uncontrolled outdoor setting is difficult, and this is one of the few times it has been attempted with this kind of rigour outside a laboratory.

Seven years on, the speakers are long gone from those trails. What remains is a dataset suggesting that a short burst of manufactured birdsong, undetected as manufactured, shifted how a group of strangers described their own state of mind on the way back to the car park.

The mosquito does not arrive as a simple machine. A female Aedes mosquito is drawn by carbon dioxide, skin odours, warmth, moisture and visual cues, but the final choice of whom to bite can also be shaped by experience. In one laboratory experiment, a human smell paired with the mechanical shock of a swat became a smell to avoid.

The finding comes from a 2018 Current Biology paper by Clément Vinauger and colleagues on host learning in Aedes aegypti, the mosquito species known for transmitting dengue, yellow fever, chikungunya and Zika viruses. The researchers trained female mosquitoes by pairing host odours with a mechanical disturbance meant to mimic the kind of defensive movement a mosquito might experience when a host tries to slap it away.

We are writers, not clinicians. This is reporting on insect behaviour and mosquito biology, not medical advice about preventing bites or disease.

The important result was not that mosquitoes dislike being disturbed. That would be easy. The important result was that the association lasted. Twenty-four hours later, mosquitoes exposed to an odour that had previously been paired with mechanical shock were less likely to fly toward that odour. In the scale of a mosquito’s life, that is not a trivial memory.

This is one set of lab experiments, not a universal rule about every mosquito encounter outdoors. But it changes the way the bite is worth imagining. The insect hovering near an ankle is not only reading the present moment. It may also be carrying a record of how a similar scent worked out before.

The old picture was already complicated

Mosquito attraction was never just smell. A 2014 Cell study by Conor McMeniman and colleagues described how carbon dioxide and other sensory cues combine to drive attraction to humans. Exhaled carbon dioxide can wake up host-seeking behaviour. Skin odours help specify the target. Heat and visual cues help guide the last part of the approach.

That layered system makes sense for a small insect trying to find blood in a noisy world. A warm object is not always a host. A smell can drift away from its source. Carbon dioxide plumes break apart in air. A mosquito has to integrate partial signals while flying through moving, changing information.

Riffell’s group had already shown that sensory cues interact in more than one direction. In a 2015 Current Biology paper, Floris van Breugel and colleagues reported that mosquitoes use vision to associate odour plumes with thermal targets. Smell can prime a visual search; heat can help confirm that the search has found something alive.

The 2018 learning paper added another layer: memory. A host’s smell is not only an attractant in the abstract. It can become part of an individual mosquito’s history.

How to teach a mosquito an unpleasant smell

The experiment used aversive conditioning, a plain idea with a technical name. Present a cue, then pair it with something unpleasant. Later, test whether the animal treats that cue differently. In this case, the cue was odour, including human-derived odour. The unpleasant event was mechanical shock, standing in for the turbulence and impact of defensive host behaviour.

After training, the mosquitoes were tested in a flight arena. The question was whether they still flew toward the trained odour or avoided it. The paper reported that females could learn to avoid odours associated with the shock and that the learned avoidance could still be detected 24 hours later.

That is a modest-sounding result until the animal is considered. A female mosquito may live only weeks. She needs blood meals to produce eggs. Host choice affects whether she feeds, survives, reproduces and possibly transmits pathogens. A 24-hour memory can cover a meaningful part of her active life.

The study did not show that a mosquito recognises a person the way a person recognises a face. Odour is not identity in the human sense. Human scent is a chemical cloud shaped by skin secretions, microbes, diet, genetics, hygiene and environment. The point is narrower and stronger: under lab conditions, mosquitoes could associate a particular host-like odour with a bad outcome and change their next choice.

Dopamine was part of the story

The title of the Current Biology paper is deliberately about modulation, not just memory. Vinauger and colleagues examined the role of dopamine, a chemical messenger involved in learning across many animals. When dopamine signalling was disrupted, the mosquitoes’ ability to learn the odour-shock association was impaired.

That matters because it ties a behaviour people notice every summer to a nervous-system mechanism. The mosquito is not choosing with a human-like mind, but neither is it a passive needle with wings. Its brain can update the value of a smell after an encounter goes badly.

Earlier work had already made mosquito learning a serious subject. In a 2014 Journal of Experimental Biology study, Vinauger, Eleanor Lutz and Jeffrey Riffell showed that Aedes aegypti could form olfactory memories in response to learned associations, including memories retained for at least 24 hours. A 2016 Trends in Parasitology review placed this in a broader disease-vector context, arguing that learning and memory can shape how blood-feeding insects respond to hosts, odours and environments.

Why one person’s odour may stop working

One of the tempting interpretations is that mosquitoes “remember people.” That is close enough to be catchy and loose enough to mislead. The study does not show a tiny social memory of a named individual. It shows learned avoidance of an odour profile after that odour was paired with an aversive experience.

Still, that is enough to make host choice more personal than the usual explanation allows. People differ in skin odour. Some are more attractive to mosquitoes than others. If an odour profile has been followed by a failed or dangerous feeding attempt, avoiding it later could be useful. A host that keeps moving, swatting or interrupting feeding may not be worth the risk when another odour is available.

There is an evolutionary logic to that. Blood feeding is hazardous. A mosquito approaching a host can be crushed before feeding, during feeding or just after. The host is food, but also a threat. Memory lets the insect treat the world as uneven: not all attractive smells are equally safe.

That does not mean swatting is a reliable repellent strategy. The lab setup was controlled, the odour presentations were defined and the mosquitoes were tested under experimental conditions. Outdoors, wind, sweat, clothing, species differences, hunger state, local mosquito density and competing hosts all complicate the picture.

A bite is a decision, not just a reflex

The value of the study is that it makes the mosquito’s choice less automatic. The insect is guided by smell and heat, yes, but also by a nervous system able to revise what those signals mean. Attraction is not fixed at the moment a plume of human scent reaches the antennae.

This is one reason mosquito biology is difficult to reduce to a single explanation. Carbon dioxide matters. Skin odour matters. Heat matters. Visual contrast matters. Hunger and reproductive state matter. The Vinauger study adds that experience matters too.

There is something unsettling in that, but also clarifying. The mosquito at the edge of the room is not thinking about you. It is not holding a grudge. But if a smell has led to trouble before, its tiny brain may have marked that odour down.

The bite begins long before the landing. It begins in a stream of cues, corrected by memory, with the mosquito deciding whether this particular scent is still worth the risk.

The two things ostriches really do are stranger, and more useful, than the legend.

A claim that cannot survive first contact with biology

The image is simple and it sticks: a frightened ostrich, head plunged into the sand, body left comically exposed. It shows up in cartoons, in sayings, and in office metaphors about denial. It is also impossible. A head packed into sand cannot draw breath, and a big bird cannot hold that pose without suffocating. National Geographic Kids puts it plainly: “Ostriches don’t bury their heads in the sand—they wouldn’t be able to breathe!”

There is a second problem. An ostrich in danger has a much better option than hiding. It is the fastest-running bird alive, reaching about 70 km/h (43 mph). A bird that can sprint that fast has little need for a hole in the ground.

What people actually watched: eggs being turned

So where did the image come from? The likeliest answer is the nest. Ostriches scrape a shallow dip into open ground, and the bird lowers its head into that dip many times a day to reach the eggs.

What the head is doing down there is careful work. The bird uses its beak to turn and reposition the eggs several times a day. Glinda Cunningham of the American Ostrich Association told National Geographic that “if you see them picking at the ground from a distance, it may look like their heads are buried in the ground.” The nest helps the effect along. These scrapes are shallow, broad depressions cleared into open ground, and a head dipped into one really does drop out of a distant viewer’s line of sight.

The second behaviour: going flat to vanish

The other source of the myth is a defence that works by disappearing rather than running. When a quick escape isn’t possible, the bird drops. The Cleveland Zoological Society describes it: “when an ostrich senses danger and cannot run away, it will flop to the ground and remain still, attempting to blend in with the terrain.” It stretches its neck flat along the earth and holds still, letting its colour blend into the dry ground.

From any real distance, a motionless ostrich pressed to the ground with its neck stretched out could look like a bird that has lost its head. The neck lies in a low line the eye struggles to pick out against sand and scrub. The illusion is close to the obvious first reading.

Why a small head on a huge body finishes the trick

Both behaviours fool us for the same reason: the ostrich’s shape. It is the largest living bird, standing over 2.8 m (9 ft) tall and weighing up to 150 kg (330 lb) in the biggest males, per Birdfact. That huge body carries a small head on a thin neck. Drop the head low, whether into a nest or flat against the ground, and the part that vanishes is tiny while the part left behind is unmistakable. The eye sees the big pale body, cannot find the head, and fills in the story it already knows.

What the myth actually records

The head-in-the-sand ostrich is a misreading that hardened into fact. Two real behaviours, turning eggs in a scraped ground nest and flattening to hide, both drop the head out of sight while the body stays fully visible. Add a small head on the largest bird alive, and a distant observer needs no cartoon to reach the wrong conclusion. The legend persists because it was an easy mistake to make. The bird tending its eggs, or lying dead-still against the ground, was doing something more considered than the myth ever gave it credit for.

The wombat fact sounds like something invented for a trivia night: an Australian marsupial produces cube-shaped faeces. The more interesting part is that the cubes are not cut, squeezed through a square opening or carved at the last moment. In a 2021 paper in Soft Matter, Patricia J. Yang and colleagues report that the corners form inside the last 17 percent of the wombat intestine, where soft material is shaped by tissue that does not behave the same way all the way around.

This is one study, not settled consensus about every detail of wombat digestion. But it is a careful attempt to answer a strangely precise biological question with anatomy, mechanical testing and a mathematical model: how can a soft tube make a shape with flat faces and corners?

The paper is about the bare-nosed wombat, Vombatus ursinus, a burrowing herbivorous marsupial from Australia. Wombats are already unusual animals in several ways: compact bodies, powerful digging limbs, slow digestion and a habit of leaving droppings in visible places. The cube-shaped scat has often been explained in terms of signalling. A cube is less likely than a sphere or pellet to roll away from a rock, log or other raised surface. The 2021 paper does not test that behavioural explanation. It asks a different question: whatever the benefit, how is the shape physically made?

The old easy answer was the wrong kind of easy. If an animal produces a square-looking object, it is tempting to imagine that the exit must be square. The study points elsewhere. The shaping is not mainly a matter of the anus stamping a cube at the end. Yang and colleagues write that the intestine itself shows non-uniform structure, with regions of increased thickness and stiffness. In their measurements, some regions of the intestinal cross-section were about twice as thick and about four times as stiff as others.

That difference matters because the intestine is not only a pipe. It is a muscular, deforming tube. As material moves through it, waves of contraction push, dry and reshape the contents. In most animals, that process produces rounded pellets, logs or loose material because the forces around the tube are roughly compatible with round shapes. In the wombat’s final intestinal stretch, the researchers propose, alternating regions of stiff and soft tissue create a pattern of uneven deformation. Some parts move and contract differently from others, producing flat faces and corners instead of a simple cylinder.

The study combined several lines of evidence. The researchers dissected wombat intestines, compared them with pig intestines, used histology to examine tissue structure, performed tensile testing to measure mechanical properties and built a numerical model of a damped elastic ring. The model is a simplification, as models always are, but it helped test whether alternating stiff and flexible regions could produce squarer shapes under repeated contractions. In the paper’s abstract, the authors say that faster contraction in the stiff regions and slower movement in the middle of the soft regions can generate the corners.

Plainly put, the wombat is not passing a finished cube through a square die. It is gradually moulding a soft, drying material while it is still inside the bowel. The final stretch of intestine acts less like a smooth hose and more like a patterned mechanical sleeve. The stiff regions help define corners; the softer regions move differently and help form the faces between them. By the time the scat exits, much of the geometry has already been made.

That also explains why the title version of the fact can be slightly misleading if it stops at “cube-shaped.” These are biological cubes, not machine-cut dice. The edges can be rounded, the faces uneven and the pieces variable. The finding is not that wombats manufacture perfect mathematical solids. It is that their intestinal mechanics can push faecal material toward a shape that is much more angular than the droppings of other known animals.

The “only known animals” part is important but should be read carefully. It means that among animals described in the biological literature and ordinary natural-history observation, wombats are the known case of cube-shaped faeces. It does not mean every species has been inspected in equal detail. Still, wombats are unusual enough that the question drew physicists, engineers and biologists into the same paper. The authors were not merely naming a curiosity. They were asking how soft matter can be shaped into flat-sided forms without rigid moulds.

The study is also a reminder that anatomy can solve engineering problems in ways that look odd only because we are used to human tools. If a factory needs a cube, it may use cutting, casting, pressing or a mould. A wombat has none of those. It has tissue, muscle contraction, drying material and time. The paper suggests that a non-uniform soft tube can do a kind of shaping that would be difficult to guess from the outside of the animal.

There are limits to the claim. The researchers worked from available wombat tissue, not from live imaging of every stage of faeces formation inside a moving animal. The mathematical model is designed to capture the mechanics of alternating stiffness, not to reproduce every feature of digestion. The result should therefore be read as a supported mechanism, not as a complete account of wombat bowel physiology. That is still a substantial step beyond the old assumption that the shape must be imposed at the exit.

What makes the result satisfying is that it replaces a cartoon explanation with a physical one. The corners are not a last-second trick. They are the outcome of repeated contractions in a tube whose walls differ around the circumference. In that sense, the wombat’s cube-shaped scat is not just a biological oddity. It is a small example of how living tissue can guide shape, one squeeze at a time, using softness and stiffness in the same structure.

I had read years of pop-psychology headlines without ever asking the one question that turns out to be the whole story: who, exactly, sat for these studies?

Not “humans,” the way the headlines implied. A remarkably specific kind of human, recruited from a remarkably specific kind of place, usually because they happened to be enrolled in an undergraduate course taught by the person running the study.

The acronym that quietly undermines a century of headlines

WEIRD stands for Western, Educated, Industrialized, Rich, Democratic. It was coined in 2010 by the psychologists Joseph Henrich, Steven Heine, and Ara Norenzayan, in a paper for Behavioral and Brain Sciences that did not bother softening its own title: “The weirdest people in the world?” Their argument wasn’t that Western researchers had been sloppy. It was that an entire discipline had spent decades quietly mistaking one small, historically unusual slice of humanity for a stand-in for the species, and had built a science of “the mind” almost entirely out of it.

The ninety-six percent figure traces back to an earlier audit by the psychologist Jeffrey Arnett, who went through six major American psychology journals and found that the overwhelming majority of published studies — from a discipline that claims to describe how people think, generally — had recruited their human subjects from Western, industrialized nations, the United States most of all. Henrich’s team picked up that finding and asked the question that actually reframes it: what happens if the group psychology studies most is also, on a long list of measurable traits, the strangest one to study?

Twelve percent standing in for everyone else

That reframing is the part that actually got to me. WEIRD populations aren’t a neutral baseline that other cultures deviate from around the edges. On a striking number of dimensions, they’re the outliers — the exception quietly generalized into the rule. Henrich and his colleagues pulled together cross-cultural data showing that Western subjects are unusually susceptible to certain visual illusions that people in other societies barely register, unusually individualistic in how they describe themselves, unusually inclined toward particular notions of fairness in economic bargaining games, unusually analytic rather than relational in how they reason about cause and effect.

Psychology spent a century describing “the human mind,” largely by asking the one demographic least representative of it.

Some of the specific examples in the original paper are the kind of detail that rearranges how you think about a whole field. The Müller-Lyer illusion — two lines of equal length that look different because of the angled fins added to their ends — is treated in introductory psychology courses as a near-universal quirk of human visual perception. It isn’t. Researchers testing populations outside industrialized, carpentered environments, places without a lot of rectangular buildings and right angles training the visual system a certain way, found the illusion barely registers for some groups at all. The leading explanation — that it’s a side effect of growing up around a lot of straight walls — is itself now disputed by researchers revisiting the original data, but the cross-cultural gap in susceptibility is well established.

The same pattern shows up in economic games designed to measure fairness. In the Ultimatum Game, one player proposes how to split a sum of money and the other can accept or reject the offer; reject, and both players get nothing. American undergraduates, the population the game was calibrated on, reliably punish offers they consider too stingy, even at a cost to themselves — a finding written up for years as evidence of an innate human sense of fairness. Run the same game across small-scale societies with different economic structures, and both the offers people make and the offers people are willing to accept shift substantially, sometimes in the opposite direction. The “innate” instinct was, at least in part, a learned one, calibrated to a market economy most of the species has never lived inside.

Realizing I qualify for that twelve percent — Western-adjacent enough, educated enough, online enough — didn’t feel like a personal discovery so much as an uncomfortable audit of everything I’d filed away as settled fact from a psychology headline. How much of “humans tend to…” was actually “people who read the same three magazines as the researcher tend to…”? I don’t think I’ve read a “landmark study finds” headline the same way since.

What I do with this now

I’m not arguing anyone should throw out a century of research, and neither did Henrich, Heine, and Norenzayan — their point wasn’t that WEIRD psychology is wrong, only that it has been badly oversold as universal. What’s changed for me is smaller and more useful than a takedown: I ask, almost automatically now, who was actually in the room before I let myself believe a finding applies to “people” in general.

Henrich later expanded the whole argument into a book, The WEIRDest People in the World, tracing how unusual historical forces may have shaped Western psychology in the first place — you can read more of his own framing of the original argument on his faculty page. It’s not a flattering label to end up wearing. It is, at least, an accurate one, and accurate turns out to matter a great deal more than flattering once you’re actually trying to understand how any of this works.

I’ve watched the documentary. I’ve read the advice about beans, walking everywhere, and going to bed when the sun does. I wanted the Blue Zones story to be true in the uncomplicated way it was sold to me: five particular places on Earth where people simply live better, and longer, than everywhere else, and all you had to do was copy the diet. It mostly isn’t true, and the way it fell apart turns out to be a far better story than the original claim ever was.

The postcard version

The term was popularized in the mid-2000s by the explorer and author Dan Buettner, working with National Geographic, to describe five regions with reportedly extraordinary concentrations of centenarians: Okinawa in Japan, Sardinia in Italy, the Nicoya Peninsula in Costa Rica, Ikaria in Greece, and the Seventh-day Adventist community of Loma Linda, California. It became a genuine global wellness franchise — bestselling books, a Netflix series, an entire aisle of branded food products — all resting on the premise that these five places held a reproducible formula for exceptional human lifespan, one you could import directly into your own kitchen if you were willing to eat enough beans.

A demographer starts checking the paperwork

Dr. Saul Justin Newman, a demographer at University College London, spent years doing something almost nobody else had bothered to do: auditing the actual age records behind the claims, region by region, rather than taking the reported statistics at face value. His conclusion, laid out in a widely circulated paper, was that the regions reporting the highest rates of “supercentenarians” — people aged 110 and older — were predicted far more reliably by poverty, poor birth-certificate coverage, and generally thin record-keeping than by anything resembling unusually good health, diet, or lifestyle.

Dead relatives, live pensions

Some of the discrepancy was simple clerical error — a birth year transcribed wrong decades ago and never corrected. Some of it was something closer to fraud, with a very mundane motive behind it: in regions with weak record-keeping, families have a direct financial incentive to keep a deceased relative listed as “alive” on paper, since pension payments often continue as long as no death gets formally registered. By Newman’s estimate, at least 72 percent of Greek centenarian records belonged to people who were actually dead, missing, or the product of an ongoing pension claim rather than an actual living 100-year-old. Tracking down roughly 80 percent of the world’s reported supercentenarians, he found irregularities running straight through the data, including the record-holder for the world’s oldest man, who was on file with three separate recorded birthdays depending on which document you checked. Fortune covered the findings in detail.

What makes Newman’s work land differently from ordinary myth-busting is the sheer tedium of the method. He wasn’t running a new study on diet or genetics. He was cross-referencing birth records, death records, pension rolls, and census data across regions that, in several cases, had never digitized their paperwork in the first place — the unglamorous, spreadsheet-level work that produces almost no headlines on its own until it eventually produces a headline that undoes twenty years of somebody else’s. It’s the kind of finding that makes you wonder how many other “remarkable” statistics are actually just a byproduct of nobody having checked the source documents closely enough to notice they don’t add up.

The highest rates of “extreme longevity” weren’t found in the places with the best health data. They were found in the places with the worst record-keeping.

An award for finding out the data was rotten

In September 2024, Newman was awarded the first-ever Ig Nobel Prize in Demography, an award with a comic reputation that is nonetheless given for real, peer-reviewed research that makes people laugh and then, ideally, think harder than they expected to about a topic they thought was settled.

The citation credited him with demonstrating that extreme-old-age data is likely dominated by clerical errors and outright fraud rather than by biology — a considerably less flattering explanation than “these people simply know something the rest of us don’t.”

What survives the debunking

Buettner’s team and Blue Zones LLC have pushed back on Newman’s conclusions, issuing their own demographers’ statement in response, and the underlying dispute over the data is still active rather than settled. I’m not in a position to referee that fight, and I don’t think most of us need to be. But even outlets covering the controversy have landed on a distinction worth keeping: the lifestyle habits Blue Zones popularized — plant-heavy diets, regular walking built into daily life rather than scheduled as exercise, strong social ties, a sense of purpose that extends past retirement age — remain independently well-supported for healthy aging, whether or not any particular village actually produces more 100-year-olds than its record books can prove. CNBC laid that distinction out clearly.

The habits were never the lie. The claim about who was actually living long enough to prove they worked was, and it took someone willing to do the extremely unglamorous work of counting birth certificates, region by region, to find that out.

I’d still like to believe in a postcard version of a long life. I’ve just learned to ask, first, whether anyone checked the paperwork behind the postcard.

It’s also wrong.

The mistake is bigger than a rounding quibble: it gets the shape of a dog’s life backward. Dogs don’t age at a steady pace. They age fastest when young and then slow down, which is why a one-year-old dog has more in common with a human teenager than with a seven-year-old child.

What I find interesting is how long the seven-year rule survived despite being so easy to disprove. In 2020, a team of researchers gave us something better to replace it with.

Where the seven-year rule came from, and why it stuck

The math behind the myth is almost embarrassingly simple. Mid-twentieth-century humans lived to around 70, dogs to around 10, and someone divided one by the other. The American Kennel Club traces the 7:1 ratio to that tidy coincidence, not to any measurement of how dogs actually develop. It stuck because it was memorable, and because most people never had a reason to check it.

But you don’t need a lab to see the problem. As Trey Ideker put it, “a nine-month-old dog can have puppies, so we already knew that the 1:7 ratio wasn’t an accurate measure of age.” By the old math, a human at the same point would be about five. The rule falls apart the moment you look at it.

Veterinarians have been saying as much for a while, and it shows in how the profession now describes a dog’s life. American Animal Hospital Association guidance breaks it into four life stages — puppy, young adult, mature adult, and senior. 

What a one-year-old dog actually looks like

By veterinary developmental milestones, a one-year-old dog is closer to a 15-year-old human than to a child. A dog at one has finished most of its growing, can reproduce, and behaves like an adolescent. Calling that “seven” undersells how much has already happened.

The old rule gets more than the ratio wrong; it gets the direction wrong too. Aging isn’t spread evenly across a dog’s life — it’s packed into the early years and then tapers off. A puppy’s first twelve months cover an enormous amount of ground, and each year after that covers a little less.

The 2020 study and the logarithmic curve

The best evidence for that came in 2020, when a team led by Tina Wang and Trey Ideker at UC San Diego, working with UC Davis and the National Human Genome Research Institute, published a study in Cell Systems. Instead of guessing from lifespans, they looked at chemical marks on DNA. As a body ages, small molecules called methyl groups attach to DNA in patterns that shift over time. Scientists call this pattern an epigenetic clock. Read the clock, and you can estimate how old a body is on the inside.

The team measured these marks in 104 Labrador retrievers, from puppies to old dogs, and compared the results to the same data in humans. When they matched one to the other, the relationship wasn’t a straight line but a curve. From that curve they built a formula, which genome.gov describes as a newer formula for comparing human to dog years: human age equals 16 times the natural logarithm of the dog’s age, plus 31.

The natural logarithm is a piece of math that rises steeply at first and then flattens out, which is exactly what the dogs’ data did. This is one study built on one dataset, so it’s a strong clue rather than the last word. But the shape of its finding matches what vets already knew about how dogs grow up.

What the formula means in practice

Run a one-year-old dog through it and you get about 31, or roughly a 30-year-old human in the press release’s rounding. That sounds high until you remember what the curve is saying: the early years count for a lot, then the pace eases. A dog isn’t adding seven human years every year forever. Most of the aging happens up front.

The honest caveat is that this formula was built entirely on Labrador Retrievers, and breeds age differently. A Great Dane and a Chihuahua don’t run on the same clock, so a single curve from a single breed can’t capture all of that. The researchers were upfront about it.

What I take from the study isn’t a precise conversion table so much as a better mental model. The old rule pictures a dog’s life as a steady march. The curve shows a fast start that slows, which is closer to what you can watch happening in front of you. Even Ideker seemed to take it that way. He said, “I have a six-year-old dog—she still runs with me, but I’m now realizing that she’s not as ‘young’ as I thought she was.” If there’s one thing to carry away, it’s a corrected intuition rather than a lookup table: your dog is probably further along than the seven-times rule suggested, and most of that distance was covered in the first couple of years.

It feels obvious. Closer to the fire, more heat. The idea is so intuitive that NASA Space Place opens its explainer by naming it directly: “Many people believe that Earth is closer to the Sun in the summer and that is why it is hotter.” 

The truth runs exactly backwards from the hunch. Earth is actually at its closest to the Sun in the dead of Northern Hemisphere winter. What really drives the seasons is a tilt most of us never think about.

What the numbers actually say about January

Every year in early January, Earth reaches its nearest point to the Sun. In 2026 that moment falls on January 3, about two weeks after the December solstice. Six months later, in July, Earth swings out to its farthest point while the Northern Hemisphere bakes through summer. If distance were the driver, the seasons for half the planet would be flipped.

The distance change is also smaller than most people imagine. Earth sits about 91.4 million miles from the Sun at its closest, versus 94.5 million miles at its farthest. That gap of roughly three million miles sounds large, but against the whole distance it amounts to only about a three percent swing. You can see how minor it is in the sky: the Sun’s apparent size changes by a little over three percent between the two points, a difference no naked eye would ever catch.

Why the tilt is doing the real work

Earth spins on an axis that is not straight up and down relative to its orbit. Most sources round this to 23.5 degrees but it drifts slowly over long stretches of time, between roughly 22 and 24.5 degrees. The tilt keeps pointing the same way as Earth circles the Sun, so across a year each hemisphere spends part of its time leaning toward the light and part leaning away.

That lean is what Seth McGowan, president of the Adirondack Sky Center in upstate New York, points to. He told National Geographic that “Earth’s 23.5-degree axial tilt is the primary driver of our weather and temperature shifts.” The distance effect, by comparison, is a rounding error. That same reporting notes the Sun’s intensity rises only about seven percent when Earth is closest, while the tilt changes the solar energy reaching the middle latitudes by roughly 50 percent, and far more than that near the poles.

The angle of light, not the distance

What the tilt changes is not how far the light travels but the angle at which it lands. When your hemisphere leans toward the Sun, the rays hit more directly, concentrating their energy on a smaller patch of ground, and the days run longer. Lean away, and the same light arrives at a shallow slant, spread across more surface, over shorter days. The angle and the length of daylight move together, and both trace back to that single lean.

The orbit does leave a faint mark. Because Earth moves faster when it is closer to the Sun, the timing of the seasons is very slightly uneven. As Jason Steffen, an assistant professor of physics at the University of Nevada, Las Vegas, put it, the orbit “makes the summer in the Southern Hemisphere slightly longer than the winter, and the winter in the Northern Hemisphere slightly shorter than the summer, but only by a few days.”

My read is that this misconception persists because it isn’t a foolish one. It follows a rule that works everywhere else in daily life: closer to the heat means warmer. The catch is that orbits don’t run on kitchen-stove logic. So the next time the Northern Hemisphere is shivering through January, remember that is exactly when Earth is actually nearest the fire. The heat was never about how close we got. It was about which way we were leaning.

We are writers, not clinicians. What follows is a reading of the research and of the distinctions clinicians draw, not medical or psychological advice, and none of it is a way to diagnose yourself or the person at the next desk.

The phrase that tends to attach to this behaviour is “a little OCD,” usually offered as a compliment. It is worth slowing down on that phrase, because it folds several different things into one, and the fold is where the confusion begins.

What the praise tends to miss

Obsessive-compulsive disorder, as clinicians describe it, is not a fondness for order.

It is a loop.

An intrusive thought or image arrives, it brings distress, and a behaviour follows that eases it for a while before the whole sequence returns. The behaviour can present as tidiness, or checking, or a need for symmetry. Seen from outside, it can look like unusually high standards. Experienced from inside, it is more often something the person would switch off if they could.

Clinical writing marks this with two terms: ego-dystonic and ego-syntonic. Most OCD obsessions are ego-dystonic, meaning they feel foreign, unwanted, and at odds with the person’s own values and sense of who they are.

That is the part the admiring colleague cannot see. The tidy desk is visible. The reason for the tidy desk, and what it would cost to leave it untidy, is not.

OCD, tidiness, and the condition it gets confused with

Here the language really matters, because two different things are being described by the same casual shorthand.

The first is ordinary orderliness. Liking a clean inbox is not a disorder, and calling it one drains the word of meaning.

The second is a separate diagnosis that sits much closer to the “disciplined” reading than OCD does. Obsessive-compulsive personality disorder, defined in the DSM-5 as a pervasive preoccupation with orderliness, perfectionism, and mental and interpersonal control, tends to be ego-syntonic. The person generally experiences their perfectionism as reasonable, even admirable, rather than as a symptom they want gone. In much of the clinical literature this is the condition whose traits actually resemble what people mean when they praise someone for being “so organised.”

The two are not the same, and they are frequently confused, including in the compliment itself. What complicates matters further is that OCD has presentations that blur the line. Perfectionism-themed OCD, and the drive for things to feel “just right,” can sit more comfortably with a person’s self-image than, say, intrusive thoughts of harm. When the compulsion happens to line up with something a culture already rewards, it becomes far easier to mistake for a virtue, and far easier for the person carrying it to keep explaining it to themselves as one.

One number, and why it needs careful handling

The figure most often cited for how long people wait comes from the Brown Longitudinal Obsessive Compulsive Study, reported by Anthony Pinto and colleagues in 2006. In that sample, individuals first received treatment on average more than seventeen years after initially experiencing obsessive-compulsive symptoms, and around eleven years after they would have met the diagnostic criteria.

Those numbers are worth taking seriously, and they need handling with care. This is one clinical sample, not a universal stopwatch that applies to every person. More recent work, including a 2021 self-report study published in PLOS One, finds gaps that are still long but somewhat shorter, and the exact figure shifts depending on how “onset” is defined and who is being surveyed. That direction is consistent across studies. Its precise duration is not fixed.

The delay is not caused by praise, and we are not claiming it is. Stigma, cost, not knowing where to turn, and simple non-recognition all play documented roles. Misdiagnosis is part of it too. In one 2013 survey, close to forty per cent of mental health professionals failed to identify OCD from case descriptions, and a 2015 follow-up put the figure among primary care physicians at just over half. Both studies, led by Kimberly Glazier, found the taboo-themed presentations were missed most often. Social approval belongs in this picture as one more thing that can quietly delay the moment of recognition, not as the engine behind it. When the behaviour is being applauded, the person has less reason, and less permission, to name it as a problem.

Why the discipline reading is so sticky

Most cultures treat visible effort as evidence of character. We reward the person who stays late, checks again, and refuses to let a detail slide, and we tend not to ask what is driving the refusal. A behaviour that reduces someone’s anxiety and also happens to produce clean spreadsheets will be read, almost every time, through the spreadsheet.

There is a structural version of this in workplaces. The colleague whose compulsions generate reliable, meticulous output is easy to depend on and easy to promote. The system gets what it wants. Whether the person is exhausted by the mechanism producing it is not a question an organisation is built to ask, and often not one the individual feels free to raise while the results keep earning approval.

This is where the compliment does its quiet work. Praise is pleasant. It is also a signal that nothing needs to change.

What the pattern can and cannot tell you

None of this means that organised people are unwell, or that liking order is a symptom, or that a reader who recognises something here has a condition. A pattern described in general terms cannot diagnose a particular life, and the distinction between a preference, a personality trait, and a disorder is exactly the thing a casual label flattens.

The clinical picture offers a more careful way to hear the compliment. “You’re so disciplined” describes an output. It says nothing about whether the routine behind the output is chosen or compelled, comfortable or quietly draining. Only the person inside it knows, and sometimes even they have been told for so long that it is a strength that they have stopped checking.

If any of this describes something that feels less like a preference and more like a demand you cannot refuse, it helps to know that OCD is treatable, and that the evidence base is real rather than speculative. Exposure and response prevention, a form of cognitive behaviour therapy, is often the first-line psychotherapy recommended and in American Psychiatric Association guidance, it has held up in randomised trials. A doctor or a qualified therapist can say far more about an individual case than any article can, and the International OCD Foundation maintains directories for finding clinicians who work with it specifically.

The useful move, if there is one, is smaller than a diagnosis. It simply means noticing that the same behaviour can mean two entirely different things, and that the applause tells you only which one is convenient to see.

The reporting comes from a single source: the Washington Post’s October 2021 account, by Jeremy Merrill and Will Oremus, of the internal documents that former employee Frances Haugen provided to Congress and the Securities and Exchange Commission. You can read the full article syndicated at the Seattle Times. It is careful reporting drawn from company memos, but it is one outlet’s read of one document set. This is one study of a company’s internal life, not settled consensus about how the platform behaved.

The logic behind the five was mundane. In 2017 Facebook was watching a decline in how much people posted and talked to each other, and the reaction emoji gave it new signals to work with. As the company reasoned in the documents, reacting with an emoji took an extra step beyond a single tap, so it was read as a stronger sign that a post had landed. More effort in, more weight out. The intention was reasonable enough. What it selected for was the problem.

The harm was downstream of the weighting

Here the story separates into two claims that often get welded together, and the difference matters.

The first claim is the weighting itself: reactions counted five times a Like. That is a design choice, made in 2017, applied across all five emoji. The second claim is what Facebook’s own data scientists found later. By 2019, per the Post, the company’s researchers had concluded that posts drawing a heavy share of angry reactions were disproportionately likely to carry misinformation, toxicity and low-quality news. An internal study that November, reported by NBC News, found that angry, haha and wow reactions turned up more often on low-quality and misleading civic and health content than on other material. The angry face carried no more weight than love or sad. The trouble was that angry-heavy content turned out to be worse content, and the algorithm was amplifying whatever drew reactions.

So the mechanism was cruder than deliberate outrage-farming. A blunt setting boosted all strong reactions, and ran straight into the fact that the posts most reliable at generating strong reactions were often the ones designed to provoke. A staffer had flagged this early. As Nieman Lab noted in its summary of the documents, one employee asked directly whether weighting reactions five times stronger than Likes would tilt the feed toward controversial content over agreeable content. Another acknowledged the concern was possible. The warning sat in the record for three years before the weighting was meaningfully changed.

Why the correction is more interesting than the myth

The compressed version of this story, the one that says Facebook cranked up anger specifically because anger keeps people scrolling, is emotionally satisfying and slightly wrong. The technology writer Mike Masnick argued at Techdirt that the coverage buried the fact that all five reactions started equal, and that the framing made a deliberate outrage machine out of what the documents describe as a cruder mistake.

He has a point worth taking, though it does not exonerate the outcome. The distinction is between malice and a certain kind of institutional carelessness. Nobody in the documents appears to have sat down to engineer rage. What happened instead is that a company optimised for a proxy, engagement measured through reactions, and the proxy quietly favoured the content least worth favouring. That is a more common failure than deliberate manipulation, and a harder one to legislate against, because everyone involved can honestly say they were only trying to measure interest.

The pattern under the pattern

This is where the Facebook case stops being about Facebook. Any system that measures a thing in order to reward it eventually rewards whatever games the measure. Charles Goodhart’s observation, later sharpened into the line that a measure which becomes a target stops being a good measure, was made about monetary policy in 1975. It keeps reappearing wherever institutions try to run themselves on a single number.

What the reaction emoji added was speed and reach. A newsroom that chases clicks distorts itself slowly, over editorial meetings and reworked headlines. A ranking algorithm applies the same distortion to billions of feeds in an afternoon, invisibly, with no one in the loop deciding that this particular divisive post about abortion or guns should go to a wider audience today. The Post reported that the angry reaction skewed heavily toward divisive topics, with one outlet, Fox News, drawing nearly double the angry reactions of any other page. No editor chose that. The weighting chose it.

The uncomfortable part is that the emotional impression the emoji were meant to capture was real. People genuinely did react more strongly to those posts. The signal was real enough. What had come apart was the link between strength of feeling and worth of content, and the system had no way to tell the difference between a post that moved someone and a post that inflamed them.

What the record does and does not settle

Facebook walked the weighting back in stages. It cut the reactions to four times a Like in 2018, built a mechanism to demote content drawing disproportionate angry reactions in 2019, dropped all reactions to one and a half times a Like in 2020, and eventually set the angry reaction to zero. A company spokesperson told the Post it continued working to understand and reduce content that produced negative experiences, including posts with a disproportionate share of angry reactions.

What the documents show is a specific sequence of internal choices at one company over roughly three years, as characterised by one newspaper reading a leaked cache. What they do not show, and cannot, is the counterfactual. We do not know how much of the outrage in public life over those years was manufactured by a ranking weight and how much was already there, waiting for any amplifier at all. The honest position is that the weighting made a bad tendency worse, at scale, and that the people who built it were warned and were slow.

The thing to watch is not whether a single reaction gets set to zero. It is whether the underlying habit changes: running enormous social systems on a proxy for attention, then discovering three years later what the proxy was actually selecting for.

The surge is being powered by American shoppers rather than the Chinese market that once defined K-beauty, and it cements Korea’s transformation from a regional trend into one of the largest beauty exporters on the planet. Yet the timing is complicated, because a new 15 percent United States tariff now hangs over the very market doing the most to drive the boom.

A record built in a single half-year

What stands out is not just the total but the pace behind it. Exports in the second quarter reached roughly $3.9 billion, up about 25.8 percent on the first quarter’s $3.1 billion, meaning the industry did not simply enjoy one strong month but accelerated as the year went on. That kind of quarter-on-quarter climb is unusual for a category as mature as cosmetics, and it points to demand that is widening rather than peaking.

The geography of that demand tells its own story. The United States bought around $1.45 billion of Korean cosmetics in the first half, roughly 20.7 percent of the total and up an eye-catching 41.5 percent year on year, cementing its place as the single largest destination. China, long the industry’s centre of gravity, slipped to about $1.01 billion, a 6.6 percent decline that dragged its share of Korean beauty exports down from 19.6 percent to 14.4 percent. Japan followed at roughly $580 million. For an industry that once treated the Chinese consumer as its growth engine, that reshuffling is nothing short of a structural shift.

How America became the biggest buyer

The American appetite did not appear overnight, but the scale of the swing is striking. The United States overtook China as Korea’s top cosmetics market for the first time last year, with exports there climbing from about $841 million in 2021 to $2.2 billion in 2025, even as sales to China roughly halved over the same stretch. A category that spent a decade leaning on duty-free counters and Chinese resellers has quietly rebuilt itself around a very different customer.

Much of that shift runs through social media and the shelves it now influences. K-beauty has become a fixture on TikTok and Instagram, where routines built around “glass skin” and gentle, ingredient-led formulas travel fast, and where nimble indie labels such as Beauty of Joseon, Anua, SKIN1004 and COSRX can test a product, ride a trend and sell direct through Amazon and TikTok Shop at a speed legacy giants struggle to match. Those same brands have moved onto mainstream American shelves at Sephora, Ulta, Walmart and Costco, while the retail chain Olive Young has pushed aggressively into the US market. Crucially, the customer base has broadened well beyond Asian shoppers, with Black and Hispanic creators helping carry Korean skincare into the beauty mainstream.

The affordability helps too. Where a prestige serum from a European house might cost $80, a well-reviewed Korean equivalent often lands closer to $15, and in a stretched economy that combination of low price and novel ingredients has proved hard to resist. American sales of Korean beauty products are now estimated at around $2.4 billion over the past year, a jump of nearly half on the year before. On Amazon, Korean products reportedly sell around three times faster than the average beauty item, a velocity that keeps them near the top of the platform’s rankings and pulls new brands in behind them.

From niche to the world’s number two

Zoom out from the half-year figure and the longer arc is even more dramatic. Across the whole of 2025, Korea exported about $11.4 billion of cosmetics, enough to overtake the United States and become the world’s second-largest cosmetics exporter, trailing only France and its roughly $24.3 billion. That pushed the country’s cosmetics trade surplus past $10 billion and extended its reach to more than 200 countries, a remarkable spread for an industry that not long ago was synonymous with a handful of Asian markets. Korea still sells barely half of what France ships, which the industry tends to frame not as a ceiling but as headroom, a measure of how much of the global market is still there to win.

The strategic lesson inside those numbers is diversification. Stung by past political friction with Beijing and by the volatility of relying on one dominant buyer, Korean brands deliberately spread their bets, chasing growth in North America and Europe and picking up momentum in newer markets from Poland to the United Arab Emirates. The first half of 2026 suggests that gamble is paying off, with no single country now able to make or break the industry’s year on its own.

The tariff cloud over the boom

The obvious threat is the one attached to K-beauty’s best customer. Under a trade deal struck with President Donald Trump, South Korean goods entering the United States now face a 15 percent tariff, a levy that lands squarely on the market responsible for more than a fifth of Korea’s cosmetics exports. For now, many brands have chosen to absorb the extra cost rather than pass it to shoppers, while others are weighing whether to manufacture or ship from inside the United States to soften the blow.

Whether that holds is the open question hanging over an otherwise triumphant year. A record $7 billion in six months shows how far K-beauty has travelled from niche curiosity to global heavyweight, but it also shows how much now rides on American shelves and American prices. If the tariff eventually feeds through to the checkout, will the shoppers who turned a Korean serum into a TikTok staple keep reaching for it, or does the next chapter of this boom depend on how much of the cost Korea is willing to swallow to stay there?

The find matters for far more than its record-breaking age. The water is not sterile; it has been quietly sustaining microbes that have never once relied on sunlight, and its chemistry is now helping scientists decide where to look for life beyond our planet, from the deep subsurface of Mars to the buried oceans of icy moons.

Reading a two-billion-year-old clock

The water comes from Kidd Creek mine, one of the deepest base-metal mines on the planet, where crews dig for copper, zinc and silver more than two kilometres down. In 2013, researchers there recovered water around 1.5 billion years old from fractures at 2.4 kilometres, already a startling result. When miners drilled deeper still, close to three kilometres, they struck an even older fluid seeping from the rock, and testing pushed its age back to roughly two billion years. The results, unveiled at the American Geophysical Union’s 2016 meeting, quickly travelled around the world and earned the fluid a place in the record books as the oldest known water on Earth.

Dating water is nothing like dating a bone or a lump of charcoal, because there is no carbon clock to read. Instead, Sherwood Lollar’s team measured the noble gases, including helium, neon, argon and xenon, that steadily accumulate inside water as it sits trapped underground, fed by the slow radioactive decay of elements in the surrounding stone. The longer the water has been sealed away, the more of these gases it collects, which let the researchers read it almost like a set of invisible tree rings and settle on an age somewhere between 1.5 and 2.6 billion years. For Sherwood Lollar, that isolation is the whole point, because it means the fluid is a sealed sample of the planet’s distant past, its chemistry preserved since an age when Earth’s atmosphere and oceans looked nothing like they do today.

Salty, syrupy and very much alive

There is no glowing underground lake here, only water leaking from cracks in ancient stone, litre by litre, and it is nothing like the fresh groundwater most of us picture. Sherwood Lollar, who has spent much of her career hunting for exactly this kind of fluid in mines across the Canadian Shield and beyond, described what she found as not “something to write home about,” explaining that the water can be “up to 10 times the salinity of sea water,” and calling it “really quite a nasty business.” That extreme saltiness is a clue in its own right, because this deep water grows saltier the longer it stays buried, so the sting on her tongue was, in effect, a rough measure of its age.

What truly sets it apart is that the water is alive. Despite two billion years without a single ray of sunlight, it hosts microbes that survive on chemistry rather than photosynthesis, feeding on the hydrogen and sulphur compounds released when water reacts with the rock around it. “To us it makes it much more interesting, it’s full of chemical energy for life, which is part of the reason why we’re investigating it,” Sherwood Lollar said. The discovery adds to a growing body of evidence that life is not merely a thin film clinging to the sunlit surface, but reaches deep into the dark, hidden parts of the planet, thriving around hydrothermal vents, in caves, and in water sealed inside solid rock. Cut off from everything else, these organisms carry on in extreme slow motion, a reminder of just how little life actually needs to keep going.

What it means for the search for life

The discovery lands in the middle of one of science’s biggest questions: where else might life exist? The guiding principle in that search has always been to follow the water, and the deep, ancient rock that held this fluid closely resembles the old crust of the Canadian Shield, which in turn resembles the ancient rock of Mars. If life can persist in isolated water far beneath Earth’s surface, then similar refuges below Mars, or inside icy moons such as Europa and Saturn’s Enceladus, which is actively spraying water into space, become far more plausible places to search. It hints that habitability may not demand sunlight, warmth or an atmosphere at all, only water, rock and time, three ingredients that turn out to be surprisingly common across the solar system.

Sherwood Lollar has made that comparison directly, noting that “one way of looking at Mars is that it looks like the Canadian Shield … in the sense that it is all ancient, billion-year-old rock on the surface,” and arguing that discoveries like this force scientists to “think outside the box about the nature of what can sustain habitability.” For decades the assumption was that life clung to the thin, sunlit skin of the planet, but water this old, this deep and this alive suggests the story runs far deeper, quite literally, than we tend to assume.

A time capsule beneath our feet

It is easy to imagine the ground underfoot as inert and finished, yet this water is proof that the planet’s depths are still active, still chemically alive, and still keeping time on a scale that makes all of human history look brand new. When this water was first sealed away, Earth’s air was only beginning to fill with oxygen and the most sophisticated life anywhere was still single-celled. It has outlasted nearly every species that has ever lived, and it surfaced only because miners happened to drill in exactly the right spot. Which raises an unsettling question worth holding onto the next time you turn on a tap: if water older than almost all life on Earth was waiting just a few kilometres down, what else is still hidden beneath us that we have not yet thought to look for?

Saudi Arabia is one of the largest desert nations on Earth. Approximately 95% of its land area is covered by sand. The country contains the Empty Quarter — the Rub’ al Khali — which is the largest continuous sand desert in the world, covering roughly 650,000 square kilometres of the Arabian Peninsula. If any nation should have an abundant, self-sufficient supply of construction sand, it should be this one.

In 2023, Saudi Arabia imported approximately $140,000 worth of construction-grade sand from Australia. The figure is small in dollar terms — Saudi Arabia imported closer to $6.45 million in construction sands overall that year, with China as the largest supplier — but the fact of the trade flow, from an Australian sand quarry to the middle of the world’s largest sand desert, captures a specific and revealing paradox about how modern construction works.

The reason Saudi sand cannot be used is not a matter of quantity, quality, or availability. It is a matter of shape.

Why wind makes sand the wrong shape

The genuinely interesting part of this story is not the trade flow. It is the specific physical mechanism by which desert sand becomes unusable for construction concrete.

When a grain of sand moves in wind, it collides with other grains — thousands of times per hour, over years, over centuries. Each collision is glancing, high-energy, and abrasive. The grain gets bounced repeatedly against its neighbours. Its sharp edges wear away first. Its faces get rounded. Over long enough periods — thousands of years is enough — the grain becomes progressively smoother, rounder, and closer to a perfect sphere. Sand geologists sometimes describe well-aged desert grains as “polished.” This is not just aesthetic. It is a specific physical transformation of the grain’s surface geometry.

Water-transported grains, by contrast, undergo a different kind of erosion. Water is much denser than air, and the movement of grains in a stream is slower, more sustained, and less collisional. Grains slide against each other or against the streambed rather than hitting each other end-on at speed. The result is that river and lake sand tends to retain more angular corners, with grains that keep some of their fracture surfaces intact. Marine sand from wave zones — where water motion is more energetic — is intermediate: smoother than river sand but still substantially angular compared to desert sand.

These are two completely different geological products of the same starting material. Under a microscope, a grain of desert sand from the Empty Quarter and a grain of river sand from the Murray-Darling system in Australia are almost unrecognisably different from each other. One looks like a small polished ball. The other looks like a tiny chunk of broken rock.

Cement, when it hardens, works by binding to the surfaces of the grains it surrounds. Angular grains offer many small flat surfaces at various angles, which the cement can grip. Rounded grains offer essentially no gripping surfaces. Smooth grains slide past one another rather than interlocking. Concrete made primarily with desert sand cracks under stress, ages poorly, and requires substantially more cement to reach the same structural strength.

For a small structure, the difference might be tolerable. For a 170-kilometre-long mirrored city being built at NEOM, or a 200-storey tower at Jeddah, or a kilometre-long bridge, it is not.

The global sand crisis

The Saudi paradox is one visible instance of a much larger and less-appreciated global phenomenon.

According to the United Nations Environment Programme, the world consumes approximately 50 to 55 billion tons of sand and gravel annually. This makes sand the most extracted solid material on Earth by mass — more than fossil fuels, more than all metals combined, more than any other natural resource by a substantial margin.

The vast majority of this sand is used in construction concrete, in road base, and in glass manufacturing. The vast majority of it, in turn, comes from a specific class of sources: rivers, lakes, glacial deposits, and marine environments where the grains have been shaped by water rather than wind. These sources are physically limited. Rivers deposit sand at specific rates over specific timescales. Marine sand accumulates only in specific coastal zones. Glacial sand is essentially a finite legacy of the last ice age.

Global sand extraction is currently exceeding the natural replenishment rate. The UNEP has described this as the “sand gap,” and has warned that construction demand alone could rise by up to 45% by 2060. The material that everyone assumes is infinite because there is so much of it on beaches and in deserts turns out to be, in the specific form required by industry, a limited and depleting resource.

The Saudi situation is the sand crisis in its most compressed form. A country with essentially unlimited sand physically present within its borders cannot use any of it for the specific industrial purpose that drives most global sand demand. The desert sand and the construction sand are, in an operational sense, entirely different materials.

What could change

There are two main alternatives to river and marine sand, and both are being actively pursued in Saudi Arabia and elsewhere.

The first is manufactured sand, produced by mechanically crushing rocks — usually granite or limestone — into angular grains of the required size and shape. Manufactured sand can substitute for river sand in most construction applications, at higher cost but with domestic supply security. Saudi Arabia has begun investing in manufactured-sand production as part of its Vision 2030 infrastructure programme, and other Gulf states are following.

The second is recycled concrete, in which concrete from demolished buildings is crushed and used as the aggregate for new concrete. This is technically feasible and increasingly practised, though the recycled material has slightly different properties from virgin sand and is generally used in less structurally demanding applications.

Neither alternative is currently at the scale required to fully replace natural sand imports. But both are growing, and the direction of the industry is toward domestic manufactured-sand supply for exactly the countries — like Saudi Arabia — where imported natural sand represents both a cost and a strategic vulnerability.

The immediate paradox will not disappear entirely. Saudi Arabia will keep importing some sand for the foreseeable future, from Australia and elsewhere, because certain specialised construction applications require specific grain properties that only natural water-eroded sand provides.

But the deeper story is not really about Saudi Arabia. It is about the underlying physics that determines what kind of sand exists where, and about a global industrial economy that has quietly built itself around a form of a common natural resource that turns out to be much less common than people assume.

The world is running short of angular sand. The Empty Quarter is full of the wrong shape.

Cleopatra is in the drawer. The Great Pyramid of Giza is in the drawer. Julius Caesar is in the drawer. So are Stonehenge, and Alexander the Great, and Ramesses, and the Trojan War, and the first pharaohs of Egypt. If you asked most people to put those things in order, most could probably do it, at least approximately. If you asked them to say how far apart they are, in specific years, almost nobody could.

This is where the compression happens. The mental drawer of ancient history is largely undifferentiated inside. Things in there are simply old. The specific gaps between them — hundreds or thousands of years — are collapsed into a single sense of long-ago-ness that treats the entire drawer as roughly one temporal unit.

The Great Pyramid of Giza breaks this compression in a specific way, and the way it breaks it is worth walking through slowly.

The pyramid and Stonehenge

The Great Pyramid was completed around 2560 BCE, during the reign of Pharaoh Khufu of the Fourth Dynasty. This is a specific date, established by decades of Egyptological work using king lists, radiocarbon dating of organic material found in the pyramid’s construction, and cross-referencing with contemporary records from surrounding civilisations.

Stonehenge, meanwhile, was built in phases across roughly fifteen hundred years. The earliest earthworks — the original ditch, the bank, the wooden posts — were begun around 3000 BCE, several centuries before the pyramid. The iconic standing stones, the sarsens and bluestones, went up around 2500 BCE, roughly contemporary with the pyramid’s construction.

But Stonehenge continued to be modified and added to for another thousand years after that. The site’s last documented construction phase — a ring of pits called the Y and Z holes, dug around the sarsen circle — was completed around 1520 BCE.

By the time those final pits were dug at Stonehenge, the Great Pyramid was already 1,040 years old.

That number is worth pausing on. Think about how far away 1,000 CE feels to you now — the world of medieval kings, William the Conqueror, the height of the Byzantine Empire, the earliest cathedrals of Europe. That is how far in the past the Great Pyramid was to the last Bronze Age Britons who touched up Stonehenge. Khufu had been dead for forty generations. The Old Kingdom of Egypt had collapsed, been rebuilt as the Middle Kingdom, collapsed again, and become the New Kingdom under a completely different dynasty of pharaohs.

The Bronze Age Britons finishing Stonehenge had no idea any of this had happened. They were just finishing their own ancient project on the other side of the continent.

The pyramid and Cleopatra

This is where the compression really breaks.

Cleopatra VII, the last active ruler of Ptolemaic Egypt, was born in 69 BCE and died in 30 BCE. When she walked through Alexandria — or when she visited the Giza plateau, as she almost certainly did — the Great Pyramid was already more than 2,500 years old.

That is roughly the gap between the birth of Christ and today.

Imagine standing next to something today that was completed a full 500 years before the birth of Christ. Something older to you than the entire span of the Christian era. Something so old that in the whole recorded history you were taught, it has just been standing there, being ancient. That is the temporal relationship Cleopatra had with the Great Pyramid.

She was not looking at a recent monument. She was not looking at something built by her great-great-grandfather’s civilisation. She was looking at a structure that predated her own historical imagination by 2,530 years.

Now consider the second half of the comparison. Cleopatra died in 30 BCE. From then until now is 2,056 years.

The gap between Cleopatra and the pyramid is larger than the gap between Cleopatra and us. She lived closer, in raw years, to the invention of the smartphone than to the completion of the Great Pyramid.

Why this is difficult to accept

The reason this fact is difficult to hold is that most historical education organises the past by narrative rather than by year.

Cleopatra is filed under “ancient Egypt” alongside Khufu. Both wore linen. Both spoke Egyptian (although Cleopatra’s Egyptian was actually a rare accomplishment for her dynasty — her Greek-speaking family had ruled Egypt for three centuries without most of her ancestors bothering to learn the local language). Both were pharaohs. Both are, for most modern readers, characters in the same broad story called “the ancient world.”

The compression is educationally efficient and historically misleading. Egyptian civilisation lasted roughly 3,000 years — longer than the entire span of time from Cleopatra to the present. The gap between Khufu and Cleopatra, measured in years, is more than the gap between Cleopatra and the space age. The people at the two ends of ancient Egypt were as temporally distant from each other as we are from the fall of Rome. But they get filed together, in most minds, because “ancient Egypt” reads as a single unit.

The pyramid comparison is one of the cleanest ways to force the compression open. Once the numbers are really absorbed, the drawer stops being flat. Time inside it starts to have shape. Cleopatra stops being contemporary with Khufu and becomes, in her own experience, someone visiting an already unimaginably ancient monument. Stonehenge stops being simultaneous with the pyramid and becomes, at its final phase, a structure being touched up by people whose ancestors had been building it while the pyramid was already a millennium old.

What still stands

The strangest thing about the whole comparison is not that the numbers are extreme. It is that the pyramid is still there.

Erosion has reduced the original height of roughly 481 feet to about 455. The polished white Tura limestone casing that once wrapped the outside was stripped centuries ago, most of it repurposed for buildings in medieval Cairo. What remains is the underlying core — the rough limestone blocks, still stacked in essentially the same configuration Khufu’s masons left them in, still holding themselves up against gravity, still visible from space.

The Great Pyramid held the record for the tallest human-made structure on Earth for approximately 3,800 years — a run of dominance not matched by any other single artefact of human construction. It was finally surpassed only in the early 14th century CE, when the central spire of Lincoln Cathedral in England rose to about 525 feet. Every fortress, every observatory, every tower, every temple built anywhere in the world between 2560 BCE and 1311 CE was shorter than something the Egyptians had finished before the invention of the wheel spread beyond the Fertile Crescent.

Cleopatra saw it. Alexander the Great, three centuries before Cleopatra, saw it. The historian Herodotus, three centuries before Alexander, saw it and wrote about it as an already-ancient wonder.

We can see it too. It has, in the honest sense, seen us coming for a long time. And it will, almost certainly, still be there long after we are gone.

They look, from close up, like potatoes or apples — rough-surfaced, mostly black, roughly fist-sized. They lie on top of a fine muddy sediment that has been undisturbed for millions of years. Around and between them, in some of the least explored habitat on the planet, live thousands of species of animals — corals, sponges, worms, small crustaceans, translucent fish — that no scientist has ever formally described.

The rocks are called polymetallic nodules. And they are, by any reasonable measure, one of the most valuable naturally occurring deposits of industrial metal anywhere on Earth.

Each nodule is approximately 25 to 30 percent manganese, 1 to 2 percent nickel, 1 to 2 percent copper, and 0.2 to 0.3 percent cobalt by weight. These metals — particularly cobalt and manganese — are essential ingredients in the batteries, electric vehicles, and renewable-energy infrastructure the world is trying to build in the coming decades. The specific region of the Pacific where they are concentrated, called the Clarion-Clipperton Zone, holds an estimated 21 billion tons of these nodules in total, containing approximately 5.95 billion tons of manganese and 0.05 billion tons of cobalt.

Those figures, for both cobalt and manganese, exceed known terrestrial reserves. The largest single deposit of these two industrial metals in the world is at the bottom of the ocean.

How the nodules formed

The specific way these nodules came to exist is worth being direct about, because it is not what most people imagine when they picture “underwater rocks.”

They did not form as normal rocks form. They were not pushed up by geological activity. They were not deposited by ancient rivers. They accumulated, atom by atom, from the surrounding seawater — dissolved metals precipitating out onto small nuclei of shell, bone, or hardened sediment on the abyssal floor.

Each nodule began as something tiny. A fragment of a shark tooth. A piece of shell. A small hard particle of something dropped from above and left on the sediment.

Around that nucleus, dissolved manganese and iron in the surrounding water — held in solution at trace concentrations — began to precipitate. Not in bursts. Slowly. So slowly that the accretion rate for a typical Clarion-Clipperton nodule is estimated at approximately one millimetre of growth per one to ten million years.

The apple-sized nodules that mining companies now want to collect are, by this measure, between five and ten million years old at their outer surface. Their inner cores may be considerably older. They have been sitting on the abyssal plain, growing at the rate of a single atomic layer per year, since long before humans existed as a species.

What lives around them

The nodules are not just economically valuable. They are also, ecologically, doing something that is not immediately obvious.

The abyssal plain of the Clarion-Clipperton Zone is almost entirely soft sediment. There is no rock. There are no cliffs. The nodules are, in significant stretches of the region, the only hard surface for hundreds of kilometres in any direction. This makes them essential habitat for any species that requires something solid to attach to.

Corals grow on them. Sponges anchor to them. Small sessile animals — creatures that spend their entire lives attached to one place — depend on nodules for that place. Around them, in the sediment itself, whole communities of tiny worms, crustaceans, and microorganisms have evolved to specific microhabitats that only exist because the nodules exist.

A 2023 checklist compiled by the Natural History Museum in London catalogued approximately 5,578 animal species observed in samples from the Clarion-Clipperton Zone. Of those, only 436 had been formally named and described by science. The remaining 5,142 species were entirely new to taxonomy — organisms that had been photographed or sampled, but not yet studied enough to be given names or placed in the tree of life.

The 90-percent-undescribed figure is the conservative reading of the data. Newer surveys, using environmental DNA sampling and improved deep-sea imaging, continue to turn up species previously unknown to biology on nearly every expedition.

The asymmetry of the extraction

The specific technical proposal of the deep-sea mining industry is to send collector vehicles to the seabed, sweep up the nodules, pipe them to surface ships, and bring them to processing facilities on land.

The extraction rate the industry is targeting is, at commercial scale, on the order of thousands to millions of tons of nodules per year. In 2022, a single feasibility test conducted by The Metals Company collected approximately 3,300 tons of nodules in the space of weeks.

The formation rate of the nodules, as noted earlier, is approximately one millimetre per one to ten million years.

That asymmetry — hours of collection versus millions of years of formation — is one of the more extreme extraction rate mismatches in modern industry. In practical terms, the nodules are being treated as if they were a mineral resource on a comparable time scale to coal or copper ore. They are not. Coal takes hundreds of millions of years to form, but coal beds are thick and continuous. Nodules take similar geological time to form, but each individual apple-sized rock is a discrete, individually-grown, sedimentary artefact that will not be replaced within any human timescale.

This is why the Clarion-Clipperton Zone is regulated by the International Seabed Authority — a UN body established under the Convention on the Law of the Sea. As of 2026, 19 exploration licences covering approximately 1.5 million square kilometres have been granted. Full commercial mining has not yet begun at scale but is projected to be authorised within the current decade if the ISA’s regulatory framework is completed.

What has not been established

Because the region is so poorly studied, the full ecological consequences of nodule removal cannot currently be predicted with confidence.

What is known, from small-scale experimental mining tests and modelling studies, is that removing the nodules destroys the microhabitat for the sessile species that depend on them. Sediment plumes generated by mining vehicles smother larger areas beyond the direct mining footprint. Dewatering discharge — the seawater and sediment waste returned to the ocean from surface ships — has been shown, in recent research published in Nature Communications, to potentially disrupt midwater food webs at depths well above the abyssal floor.

What is not known is what the aggregate consequences of full-scale commercial mining across hundreds of thousands of square kilometres would be for a habitat where 90 percent of the species have not been described. It is not currently possible to conduct rigorous impact assessment on species that have not yet been named.

The nodules will still be there whether they are mined this decade or not. The species living around them, if they are lost, will not be recoverable. The specific extraction currently under consideration is genuinely, in the honest sense, one of the largest interventions in an unstudied ecosystem in the history of human industry.

Whether it happens, and at what scale, will be decided within the next few years — by regulatory bodies, corporate boards, and national governments — in decisions that will affect a place where almost no one has ever been.

On 22 December 1938, the South African trawler Nerine returned to East London with a large blue fish caught near the mouth of the Chalumna River. It was about 1.5 metres long, heavily scaled and unlike anything the museum curator called to inspect the catch had seen before.

Marjorie Courtenay-Latimer recognized that the fish was unusual and saved it for examination. The animal had been brought up alive, but it did not survive the temperature and pressure change. Ichthyologist J. L. B. Smith later identified it as a coelacanth, a group then known only from fossils, and named the species Latimeria chalumnae in her honor and for the waters where it was caught.

The event is often described as a fish returning from extinction. More precisely, a lineage that had vanished from the known fossil record turned out to have living representatives. The modern fish was not the same species found beside dinosaur bones, and its ancestors had not spent 66 million years frozen in evolutionary time.

A museum curator knew the catch was different

Courtenay-Latimer was 31 and working at the East London Museum when she examined the trawler’s haul. The coelacanth’s thick scales, paired fleshy fins and unusual tail made it difficult to place among familiar fishes. A Ditsong Museums account dates her recognition to 22 December 1938.

Smith was away when she tried to contact him, so she sent a description and drawing. By the time he examined the preserved remains, he was convinced that the fish belonged to the ancient coelacanth group. His short 1939 report in Nature, titled “A living fish of Mesozoic type”, introduced the specimen to zoology.

The surprise was not that a particular named fossil species had reappeared. The living genus Latimeria was new. What returned was an entire branch of lobe-finned fishes that paleontologists believed had ended near the close of the Cretaceous.

The fossil record stopped 66 million years ago

Coelacanth fossils extend back more than 400 million years. The Natural History Museum notes that known specimens range from about 409 million to 66 million years old. The group was once diverse, with species living in marine and freshwater environments and bodies quite unlike the modern deep-water form.

Then the record ended around the same boundary that marks the extinction of non-avian dinosaurs. With no younger fossils and no specimens in scientific collections, extinction was a reasonable inference.

It was still an inference from absence.

Fossilization is uneven, deep rocky habitats are difficult to sample, and a rare fish can leave no remains in the places paleontologists happen to examine. The 1938 catch exposed the gap between “not in the record” and “not alive.”

There is another qualification. Western science had lost the coelacanth, but fishers in the Comoros knew it. The Natural History Museum records that local people called it gombessa and occasionally caught it during night fishing. Its scientific rediscovery was not humanity’s first encounter with the animal.

Why it looked like a fish from another age

Coelacanths are lobe-finned fishes. Their paired fins contain fleshy, muscular bases supported by internal bones, unlike the thin fin rays of most living fish. As they swim, the paired fins move in an alternating pattern that can recall the limb sequence of a walking animal.

That resemblance once encouraged the idea that the coelacanth stood close to the direct ancestry of land vertebrates. Genome-scale comparisons now place lungfishes closer to tetrapods. The coelacanth is a surviving cousin from the broader lobe-finned branch, not a stranded ancestor waiting to walk ashore.

Other features are genuinely unusual. The skull contains an intracranial joint that allows the front portion to lift during feeding, and the snout houses an electrosensory rostral organ. Its tail has three visible lobes. The Smithsonian’s overview explains that living Latimeria is distinct from fossil forms even though coelacanths share a recognizable body plan.

“Living fossil” does not mean unchanged

The phrase “living fossil” is useful when it means a living member of an old lineage that was long known only from fossils.

It becomes misleading when it implies that evolution stopped. Living populations still accumulate genetic variation and remain subject to natural selection, even when their broad body shape changes slowly.

In a 2013 BioEssays review, Nicolas Casane and Patrick Laurenti argued that molecular and anatomical evidence does not support treating coelacanths as evolutionarily static. Fossil coelacanths changed shape across time, and the genomes of living species continue to mutate, recombine and respond to selection.

The 66-million-year figure therefore describes the gap in the recognized fossil record, not the age of the modern species. Latimeria chalumnae has its own evolutionary history within a much older group.

Two living species occupy scattered deep habitats

After 1938, searches found West Indian Ocean coelacanths around the Comoros and later off Mozambique, Madagascar, Kenya, Tanzania and South Africa. They are generally associated with steep rocky slopes, submarine canyons, caves and overhangs. Individuals rest in caves by day and emerge at night, moving slowly while searching for fish and squid.

They do not live only at one extreme depth. A 2020 report in the South African Journal of Science, led by Michael Fraser, summarized records ranging from about 54 metres to more than 800 metres, with temperature and oxygen helping determine where suitable habitat occurs.

A second living species was recognized after coelacanths were found near Sulawesi, Indonesia, in 1997 and 1998. The Smithsonian records the two as the blue West Indian Ocean coelacanth and the browner Indonesian coelacanth, separated by thousands of kilometres.

A long life makes accidental catches costly

Coelacanth biology is slow even by deep-sea standards. A 2021 Current Biology study led by Kélig Mahé examined growth marks on scales from 27 specimens. The team estimated a lifespan near 100 years, maturity between 40 and 69 years and gestation around five years.

Those estimates depend on interpreting extremely fine annual structures in scales, and the sample was necessarily small. Even so, the combination of slow growth, late maturity and few young means a population may replace animals lost to fishing very slowly. Coelacanths are usually caught as bycatch rather than targeted, but rarity does not make each accidental capture biologically unimportant.

The net hauled aboard in 1938 did not bring an unchanged Cretaceous fish into the modern world. It revealed that the fossil record had stopped while the lineage continued. Research since then has shifted from asking whether coelacanths survived to finding where their small populations live and how slowly they reproduce.

Under the sand and rock of North Africa, water fills porous layers of sandstone across areas larger than many countries. Some of it entered the ground when rain crossed what is now the Sahara, then moved so slowly through the rock that it remained there through tens of thousands of dry years.

Calling this “an ocean’s worth” captures the scale, but not the geology. There is no single underground sea beneath the desert. The Sahara overlies several aquifer systems, and their water occupies pores and fractures in rock rather than a vast open chamber. The total is immense, but it is not comparable to the volume of Earth’s oceans.

The age is also more complicated than one lost green era. Different samples record different recharge episodes. Some are tens of thousands of years old; water from the deepest parts of the Nubian Aquifer has been dated to hundreds of thousands of years, in places approaching one million.

The reservoir is a layered archive of repeated wet periods.

The largest stores lie under North Africa

A continent-wide assessment published in Environmental Research Letters in 2012 estimated that Africa contains about 660,000 cubic kilometers of groundwater, with an uncertainty range from 360,000 to 1.75 million cubic kilometers. The study, led by Alan MacDonald of the British Geological Survey, found that the largest volumes occur in sedimentary aquifers beneath Libya, Algeria, Egypt and Sudan.

That figure covers all of Africa, not only the Sahara, and it estimates water held in the ground rather than water that could be recovered. Both distinctions matter. A map of storage is not a map of productive wells.

The best-known Saharan reserve is the Nubian Sandstone Aquifer System. It extends beneath Chad, Egypt, Libya and Sudan across roughly two million square kilometers. The International Atomic Energy Agency calls it the world’s largest known fossil-water aquifer system.

It is not the Sahara’s only major aquifer. Farther west, the North Western Sahara Aquifer System crosses Algeria, Tunisia and Libya. Other large basins sit beneath the western and central desert. Boundaries drawn on maps simplify formations that vary in depth, thickness, salinity and permeability.

The water is inside the rock

The word “aquifer” can suggest a buried lake.

Most groundwater looks nothing like that.

As the US Geological Survey explains, groundwater fills connected pores and fractures in sand, gravel and rock, much as water fills a sponge. Sandstone can preserve spaces between ancient sand grains. If those spaces connect, water can move through the formation and reach a well.

Porosity tells geologists how much empty space a rock contains. Permeability describes how readily water can move between those spaces. A thick layer may hold a large volume but release it too slowly for a high-yield well. Clay and shale layers can trap water under pressure, while faults can either open routes for flow or act as barriers.

This is why stored volume cannot be converted directly into a promise of supply. MacDonald’s team found that high-yielding boreholes were much less widespread than total storage might imply. Water may also be too deep, too salty, too remote or too expensive to pump.

Rain entered during several greener Saharas

North Africa has repeatedly shifted between arid and humid conditions. Changes in Earth’s orbit altered Northern Hemisphere summer sunlight, strengthening the African monsoon during favorable intervals and moving its rain belt north.

During these humid periods, water ran through river networks and soaked into exposed sandstone. A 2015 Nature Communications paper led by Charlotte Skonieczny used satellite radar to identify a 520-kilometer section of a buried river system in western Sahara. The authors concluded that the larger drainage system was reactivated during several humid episodes over the past 245,000 years.

The most recent African Humid Period lasted from roughly 14,700 to 5,500 years ago. Pollen and archaeological records indicate that woodland and grassland communities extended far north of their present ranges, supporting lakes, fish, crocodiles, elephants and human settlements.

“Forested” should not be read as a wall of rainforest from the Atlantic to the Nile. The green Sahara was a changing mosaic of grassland, wooded savanna, wetlands, rivers and lakes. Earlier humid phases could be wetter or differently distributed. The water below today’s desert accumulated during more than one of them.

Radioactive atoms work as clocks

Groundwater age means the time since rain entered the aquifer and stopped exchanging freely with the atmosphere. Hydrologists estimate it from dissolved chemicals and rare radioactive isotopes whose decay rates are known.

Radiocarbon is useful over tens of thousands of years. A 2020 study in Water led by Mustafa El-Rawy, for example, summarized Nubian groundwater age estimates of about 20,000 to 49,000 years. Dating very old water requires clocks that run longer.

In a 2004 Geophysical Research Letters paper, Neil Sturchio and colleagues measured krypton-81 and chlorine-36 in deep Nubian Aquifer samples. Their results showed that some groundwater had resided below the Sahara for up to about one million years. Krypton-81 is especially useful because it has a half-life of about 229,000 years and is distributed through the atmosphere before rain carries it underground.

These dates do not mean every liter in the aquifer is equally ancient. Water mixes, flow paths differ, and some margins receive limited modern recharge. Age generally increases along long regional paths away from recharge areas.

A huge reserve can still be depleted

Fossil groundwater is not renewable on the timescale of a farm, city or government. In the hyper-arid interior, present rainfall replaces little of what large wells remove. Pumping there is closer to mining a finite reserve than drawing from a river that returns each season.

The IAEA says growing demand and declining availability from other sources have placed the Nubian system under pressure. Chad, Egypt, Libya and Sudan have therefore developed a joint management framework because pumping in one country can affect a shared aquifer.

Even a vast system can experience local drawdown around well fields. Falling pressure can increase pumping costs, dry shallower wells and change the movement of saline water. The sheer total volume says little about how long a particular community’s borehole will remain usable.

The water beneath the Sahara is real, ancient and large enough to matter across national borders. What remains uncertain is how much can be recovered in each place, at what quality and cost, without shifting the burden of depletion to neighboring users or later generations.

Space does not begin at a clean edge where Earth’s air abruptly stops. The atmosphere thins continuously, molecule by molecule, until its outermost hydrogen becomes almost indistinguishable from the surrounding interplanetary environment.

Measurements from the ESA-NASA Solar and Heliospheric Observatory, or SOHO, indicate that this faint hydrogen envelope reaches at least 630,000 kilometres from Earth, well beyond the Moon’s average distance of about 384,400 kilometres. In that narrow physical sense, every Apollo astronaut who walked on the Moon was still inside Earth’s outermost atmosphere.

The statement is accurate but easy to misread. Lunar astronauts were in a hard vacuum, not in air. They needed pressure suits and life support exactly as space travellers do. The “atmosphere” surrounding them consisted of an extremely sparse population of neutral hydrogen atoms known through its ultraviolet glow as the geocorona.

The 630,000-kilometre estimate comes from a 2019 paper in the Journal of Geophysical Research: Space Physics led by Igor Baliukin. It is a result from one reanalysis of SOHO data, not a newly discovered solid border. The exosphere has no sharp outer wall, so any quoted extent depends on the sensitivity of the instrument and on how the surrounding hydrogen background is modelled.

The atmosphere fades rather than ends

Near sea level, gravity holds nitrogen, oxygen and other gases densely enough for us to breathe. With altitude, pressure and density fall. The troposphere gives way to the stratosphere, mesosphere and thermosphere, but these named layers describe changing physical conditions, not sealed compartments.

The exosphere is the outermost recognised layer. Here collisions between particles become so rare that individual atoms can travel immense distances along ballistic paths. Some remain gravitationally bound and fall back. Others gain enough energy to escape Earth altogether.

Hydrogen dominates the far exosphere because it is the lightest element. Solar ultraviolet radiation strikes these atoms, which absorb and re-emit light at a characteristic wavelength called Lyman-alpha. Seen with the right ultraviolet instrument, that scattered light forms Earth’s geocorona, literally its glowing crown.

The geocorona is invisible to human eyes. A conventional photograph from lunar distance shows Earth as a bright blue planet surrounded by a thin optical limb, not by an enormous luminous bubble. Detecting the outer hydrogen requires ultraviolet observations and careful subtraction of sunlight scattered by hydrogen elsewhere in the solar system.

SOHO was looking past Earth for another reason

SOHO was launched in 1995 to study the Sun. It operates near the Sun-Earth L1 point, roughly 1.5 million kilometres sunward of Earth, where it can maintain a nearly continuous view of our star. One of its instruments, SWAN, maps hydrogen throughout the solar system by recording Lyman-alpha emission.

Baliukin and colleagues returned to SWAN observations made in 1996, 1997 and 1998, when the geometry allowed the instrument to map Earth’s hydrogen cloud from outside. Their analysis followed the geocorona well past the lunar orbit. An ESA account of the work reported a detectable extent of about 630,000 kilometres, close to 50 Earth diameters.

The cloud was not uniform. Solar radiation pressure pushes hydrogen away from the Sun, compressing the dayside and stretching the distribution into a longer nightside tail. The density falls rapidly with distance. Around the Moon’s orbit, ESA reported only about 0.2 hydrogen atoms per cubic centimetre on the nightside and roughly 70 per cubic centimetre on the dayside.

Those numbers explain how both statements can be true: the Moon is inside Earth’s extended atmosphere, and the Moon is surrounded by vacuum. Sea-level air contains roughly 10¹⁹ molecules in each cubic centimetre. Even the denser dayside geocorona at lunar distance is thinner by many orders of magnitude.

Apollo 16 photographed the halo from inside it

There is a useful historical twist. In April 1972, Apollo 16 astronauts John Young and Charles Duke placed a gold-plated far-ultraviolet camera in the shadow of their lunar module. Designed by physicist George Carruthers, it became the first astronomical observatory operated from another world.

The instrument photographed Earth in ultraviolet light and recorded its geocorona. NASA’s archive describes how the camera captured Earth from the lunar surface, along with emissions from atomic hydrogen, oxygen and molecular nitrogen. The observers were standing within the outer halo they were trying to image.

That vantage point was still too close to show the entire structure. The modern Carruthers Geocorona Observatory, named for the Apollo instrument’s creator, launched in September 2025 and operates near L1. From there it can view the Earth-Moon system and survey the geocorona on a scale the lunar camera could not fit into one frame.

NASA released the mission’s first-light images in December 2025. They showed Earth and the Moon in far ultraviolet, with the geocorona appearing as a fuzzy Lyman-alpha halo. Those early images primarily confirmed that the instruments were working. The mission’s broader purpose is to map how the exosphere changes with solar activity and space weather.

Why astronauts were still unquestionably in space

There is no contradiction between being inside the geocorona and being in space. “Space” is an operational and legal category, while “atmosphere” is a physical distribution of particles. Nature does not provide a line at which the last air molecule disappears.

The commonly used Karman line places the start of space at 100 kilometres above sea level. The Federation Aeronautique Internationale uses that 100-kilometre boundary for record-keeping. It is a convention based on flight and orbital considerations, not the altitude of the atmosphere’s final atom.

Earth-orbiting satellites operate above the Karman line while still encountering traces of atmosphere. At lower orbital altitudes, that residual gas creates measurable drag. The International Space Station must periodically raise its orbit because even at roughly 400 kilometres, the thermosphere is dense enough to slow it over time.

At lunar distance, the situation is far more extreme. The geocoronal hydrogen is so sparse that it provides no usable pressure, oxygen, warmth or shielding. Apollo crews faced vacuum, radiation and temperature conditions associated with deep space. Calling them “technically inside the atmosphere” describes the origin of a handful of surrounding atoms, not the environment their spacecraft had to survive.

The Moon has an exosphere of its own

The terminology becomes even more layered because the Moon also carries a tenuous exosphere. It contains atoms and molecules released from the surface by solar radiation, micrometeoroid impacts and the solar wind. NASA’s LADEE mission measured gases including helium, neon and argon around the Moon.

The Lunar Atmosphere and Dust Environment Explorer treated this as an atmosphere despite its extraordinarily low density. An Apollo astronaut was therefore simultaneously within the Moon’s own exosphere, Earth’s much larger hydrogen exosphere and the wider flow of particles through interplanetary space.

Atmospheres are not exclusive territories. They overlap with the solar wind, planetary magnetic fields and the exospheres of nearby bodies. The outermost layers are better understood as populations of particles with changing densities and trajectories than as shells drawn around planets.

Why the distant hydrogen matters

Earth slowly loses hydrogen to space. Mapping the geocorona helps researchers estimate that escape and understand how sunlight and geomagnetic activity alter the upper atmosphere. Those processes are part of the long-term evolution of planetary atmospheres and water inventories.

The geocorona can also interfere with ultraviolet astronomy. A telescope searching for hydrogen around another object must look through or near Earth’s own extended glow. Better maps allow astronomers to separate a distant signal from the foreground produced by our planet.

For lunar exploration, individual geocoronal atoms present little direct hazard. Their importance is conceptual and scientific. The Moon does not orbit beyond a simple atmospheric boundary. It moves through the faintest outskirts of Earth’s escaping hydrogen, a region shaped by gravity, sunlight and the solar wind.

So the headline is technically right, provided “inside the atmosphere” is not allowed to imply “inside air.” Earth’s familiar atmosphere becomes space long before every Earth-origin hydrogen atom is gone. Apollo astronauts crossed the accepted boundary of space within minutes of launch, travelled through near-vacuum and stood on another world, all while remaining inside a halo so thin that it took ultraviolet instruments and decades of analysis to measure its full reach.

A dog looks at its owner. The owner looks back. In a widely cited experiment, the longer that gaze lasted, the more oxytocin appeared in urine samples from both members of the pair.

Oxytocin is involved in birth, lactation, social recognition and attachment. It also participates in the feedback between human mothers and infants. The dog study therefore offered a striking possibility: domestication may have allowed two species to recruit part of the same bonding system through eye contact.

That is a meaningful result. It is not evidence that a dog experiences love in precisely the same way as a human infant, or that one molecule contains the whole biology of either relationship. The researchers measured urinary oxytocin in small groups of dogs and owners. They did not measure love, and they did not directly measure oxytocin inside the brain.

This is one study, not settled consensus. Later work supports a role for oxytocin in dog-human affiliation, but it also shows that the hormone’s effects depend on sex, breed, prior experience, relationship and social context.

What the 2015 experiment actually did

Miho Nagasawa and colleagues at Azabu University published the central paper in Science in 2015. The study tested an “oxytocin-gaze positive loop” between dogs and their owners in two related experiments.

In the first, 30 dog-owner pairs spent 30 minutes interacting in a room. Owners could talk to and touch their dogs, and the researchers measured how long the dogs gazed at them. Urine samples were collected from both dogs and owners before and after the session.

The researchers divided the dogs according to gaze duration. There were 21 dogs in the long-gaze group and nine in the short-gaze group. Only owners in the long-gaze group showed a statistically significant rise in urinary oxytocin after the interaction. Dogs in that group also showed a larger change than short-gaze dogs.

Longer dog-to-owner gaze predicted more change in both the owner’s and dog’s urinary oxytocin after the researchers included talking and touching in their statistical model. Touch showed a weaker trend. This was an association inside an interaction that included several behaviours, not proof that gaze alone caused every hormonal change.

The team also tested 11 hand-raised wolves with their human caregivers. The wolves rarely gazed into the caregivers’ eyes, and gaze did not predict an oxytocin change. The comparison supported the idea that human-directed mutual gaze became unusually important during dog domestication.

It did not cleanly separate genes from experience. Pet dogs and hand-raised wolves differ in species history, rearing conditions, daily environment and the meaning eye contact carries in their social behaviour. The authors examined early human contact and argued that it did not explain their findings, but 11 wolves cannot settle the evolutionary history of the mechanism.

The second experiment tried to close the loop

An association between gaze and oxytocin could run in either direction. Affectionate interaction might increase oxytocin and make gaze more likely, or gaze might help raise oxytocin and encourage further interaction. The second experiment tried to manipulate one side of that cycle.

The researchers gave 27 dogs either intranasal oxytocin or saline on different occasions. Each dog then entered a room containing its owner and two unfamiliar people. The humans were told not to initiate touching or talking, limiting some of the cues that could blur the effect.

Female dogs given oxytocin gazed longer at their owners than they did after saline. Their owners, who received no oxytocin themselves, subsequently showed increased urinary oxytocin. The same pattern was not found in male dogs. The researchers proposed that sex differences in oxytocin and vasopressin systems, or vigilance towards strangers, might have contributed.

That sex-specific result matters. The experiment did not show that spraying oxytocin makes every dog gaze lovingly at a person. The paper itself described the sample as small and the social effects as complicated by sex and context.

Other experiments have produced related but not identical effects. A 2017 study found that intranasal oxytocin increased owner-directed gaze in ancient Japanese dog breeds, while a 2020 experiment with retrievers tested gaze during a food-learning task and found effects that depended on the measure and condition. Oxytocin is part of the system, not a universal gaze switch.

Why researchers compared dogs with mothers and infants

Mutual gaze is a major cue in early human relationships. A mother sees an infant attending to her, responds with voice, touch and facial expression, and the infant responds in turn. Oxytocin is among the systems that help coordinate maternal behaviour, stress regulation and social reward.

Earlier human work found that maternal oxytocin responses predicted mother-to-infant gaze. Nagasawa’s group did not put mothers, infants, owners and dogs through one matched experiment. They compared the structure suggested by the dog-owner results with a feedback loop already proposed in the parent-infant literature.

The similarity is therefore at the level of a component mechanism. In both relationships, gaze can act as an affiliative signal, social interaction can engage oxytocin, and that chemistry can influence further attention or care. The relationships remain biologically and psychologically different in obvious ways.

A human infant depends on a caregiver for survival and develops through pregnancy, birth, feeding and years of maturation. A dog-owner bond is interspecies, begins under highly variable circumstances and includes learning, food, touch, play, routine and shared environment. Finding oxytocin in both does not collapse them into one relationship.

Oxytocin is not a vial of love

Calling oxytocin the “love hormone” is convenient and misleading. Oxytocin is a peptide made in the hypothalamus. It acts as a hormone in the body and as a signalling molecule in the nervous system. Its effects vary with the brain circuit, receptor distribution, individual and situation.

Modern reviews describe oxytocin as helping regulate the detection, evaluation and reinforcement of social cues. Increased social attention can support affiliation in a familiar, safe interaction. In another setting, the same increased attention may sharpen vigilance, group preference or defensive behaviour.

This is why the presence of oxytocin cannot be translated directly into a feeling. It is involved in maternal care and pair bonding, but also in birth contractions, milk release, stress responses, memory and context-dependent social behaviour. Love is an experience and a relationship, not the concentration of a single chemical.

The measurement also deserves care. Nagasawa’s team measured oxytocin in urine before and after a 30-minute session. Urine offers a practical, non-invasive sample from dogs, but peripheral oxytocin is not a direct reading of moment-to-moment signalling in the brain. A systematic review and meta-analysis of endogenous oxytocin during human social interaction found substantial differences among study designs and measurement methods.

The dog experiment’s result is best read as evidence that an oxytocin-related physiological response accompanied certain owner-dog interactions. “The same brain chemical” is broadly true because the same molecule participates in both species. “The very same biology” is too broad if it implies identical neural activity, attachment or subjective emotion.

Dogs have other ways of showing attachment

Whether dogs love people cannot be decided by oxytocin alone. Behavioural research asks related but more observable questions. Does a dog seek its caregiver when uncertain? Does the person’s presence reduce stress? Does the dog prefer familiar people, return after separation and coordinate attention or action with them?

Dogs often use human faces and gaze in ways that wolves do not, particularly when seeking information or help. Work comparing dogs and wolves in an unsolvable task found that domestication and experience both shape “looking back” towards humans. Eye contact is one part of a learned and evolved communication system.

Touch and broader social interaction also affect the chemistry. A 2014 experiment reported that intranasal oxytocin increased dogs’ affiliation towards owners and dog partners, with further changes in endogenous oxytocin. That result fits a social feedback account extending beyond eye contact alone.

A 2019 review of the dog-owner literature nevertheless urged caution. It noted variation across studies and raised questions about whether oxytocin changes are specific to owners, familiar humans or positive social contact more generally. The review concluded that the dog-owner oxytocin story remained more complex than a single mutual-gaze pathway.

What the shared chemistry does tell us

The 2015 study remains important because it described a plausible route by which domestication could bridge two species. Dogs that comfortably held human gaze may have elicited more attention and care. Humans responsive to those cues may have formed stronger bonds with those dogs. Across generations and within individual relationships, behaviour and physiology could reinforce one another.

That loop need not imitate infancy in every detail to be significant. Evolution commonly reuses existing systems. A pathway involved in attachment within a species can be recruited into a relationship across species when both partners reliably exchange the right cues.

Owners also need not wait for a hormone assay to recognise a dog’s attachment. The animal’s choices, proximity, excitement at reunion, willingness to seek comfort and patterns of shared attention are evidence at the level where relationships actually happen. Oxytocin helps explain how some of those interactions may become rewarding and self-reinforcing.

So the headline contains a real biological insight wrapped in an overstatement. Dog-owner gaze can engage the same oxytocin system that participates in mother-infant bonding. That does not mean every gaze releases it, that the response is identical in every dog, or that canine love has been chemically proved.

What the research supports is subtler: humans and dogs appear able to enter a shared physiological feedback loop through social attention. The molecule is ancient. The interspecies relationship built around it is the unusual part.