Unstable Gas, Ancient Mars, Sinking Cities · Morning Gist
Saturday, October 315 stories · 16 min
Unstable Gas, Ancient Mars, Sinking Cities
New maps of interstellar gas, a fly's taste circuits and sinking cities test old assumptions, while an ancient meteorite and a sideways jet fill in rare details.
Picked from 53 new articles across 10 sources · arrow keys for other days
Today in 30 seconds
30s left
A Technion 3D map finds 41% of nearby interstellar gas is "unstable", challenging a textbook model. This is a preprint, not yet peer-reviewed. story 1 ↓
A fruit fly wiring map shows sweet signals from legs and tongue reach the same circuits, and taste works by lifting a constant brake. story 2 ↓
Saharan meteorite Teghaza 001 is at least 4.1 billion years old, only the third known piece of Mars's ancient crust. story 3 ↓
GRB 260310A started with a faint flash but had a huge afterglow. Astronomers think it was a powerful jet seen from the side. story 4 ↓
Satellite radar finds sinking ground in 465 big cities. For almost half of them, nobody had measured this from space before. story 5 ↓
AI on weather radar counted 156 million migrating birds in one spring night, 20 million more than forecast.
1/5Quanta Magazine· Perri Thaler· Oct 30
New 3D map finds 41% of nearby interstellar gas in an "unstable" state
Researchers at the Technion in Israel combined 3D maps of dust and starlight around the sun. They found that 41% of the nearby gas between stars sits at temperatures textbooks say it can't hold for long.
Why it matters · If the result holds, it could explain why galaxies turn so little of their gas into stars. It also shows how new 3D mapping can test models astronomers have long accepted.
Know this first
Context from the editor, not the article · 32 sec
1 · The interstellar medium
Space between stars isn't empty. It holds a thin fog of mostly hydrogen and some helium, plus dust, at roughly 130 atoms per liter near the solar system. Stars are born from this material and return gas to it when they die.
2 · The two-phase textbook model
Since 1969, the standard picture has held that this gas settles quickly into one of two stable states: cold clouds near 100 kelvins or warm diffuse gas near 8,000 kelvins. Gas at in-between temperatures was expected to be rare.
3 · Why it was hard to see
Cold hydrogen gives off no visible light. Radio telescopes detect its faint 21-centimeter signal, but every cloud along a line of sight blends into one signal. That made it hard to locate individual clouds in 3D or measure their temperatures.
2/5The Transmitter· Lauren Schenkman· Oct 30
Fruit fly wiring map traces taste from body to brain to behavior
A team at the Champalimaud Centre identified every taste neuron on fruit flies and used existing whole-brain wiring maps to follow them into circuits for walking, swallowing and insulin release. The work appeared in Cell.
Why it matters · Researchers can now switch individual fly neurons on or off along a known circuit and test exactly how a sensation becomes an action, which is rarely possible.
Know this first
Context from the editor, not the article · 27 sec
1 · What a connectome is
A connectome is a complete map of the neurons in a nervous system and the synapses joining them. It is usually built by slicing tissue, imaging it with electron microscopes, and tracing every cell, often with AI help.
2 · Fly connectomes so far
Fruit flies (Drosophila) are the first complex animals with near-complete wiring maps: a whole female brain, released in 2024, and now a full male nervous system. Each took years of imaging and tracing work.
3 · Flies taste all over
Fruit flies have taste-sensing neurons on the tongue-like proboscis, in the throat, on the legs and on the wings. These neurons detect sweet, bitter, salt, water, carbonation, fats and amino acids.
3/5Ars Technica Science·
Jacek Krywko
· Oct 30
Saharan meteorite is a third known piece of Mars's ancient crust
A NASA-led team showed that Teghaza 001, a Martian meteorite found in Mali in 2022, is at least 4.1 billion years old. Its chemistry points to active water in early Mars's crust and an unusual deep source.
Why it matters · Most Martian meteorites on Earth are much younger than the planet's surface, so a single intact ancient rock is rare direct evidence of what early Mars was like.
Know this first
Context from the editor, not the article · 29 sec
1 · How Mars rocks reach Earth
A big asteroid impact on Mars can blast rock fast enough to escape its gravity. Some of those fragments eventually fall on Earth as meteorites. Deserts and Antarctica are good places to find them, because dark stones stand out there.
2 · Dating surfaces versus rocks
Scientists estimate the age of a planet's surface by counting craters: older ground has piled up more impacts. A rock in the lab can be dated directly instead, by measuring how far radioactive elements in its minerals have decayed.
3 · Isotopes as fingerprints
Isotopes are forms of an element that differ in mass. Each planetary body has its own characteristic mix of oxygen isotopes, so measuring a meteorite's oxygen can show which world it came from.
4/5Universe Today· Evan Gough· Oct 30
A weak gamma-ray burst turns out to be a powerful jet seen from the side
Astronomers studied a gamma-ray burst that looked weak at first, then grew unusually bright. A peer-reviewed study argues that we saw its two-layered jet from off to the side, not head-on.
Why it matters · Most gamma-ray burst models are built from jets aimed straight at us. A burst seen from the side shows the jet's structure and how viewing angle can bias brightness estimates.
Know this first
Context from the editor, not the article · 32 sec
1 · What a gamma-ray burst is
Long gamma-ray bursts happen when the core of a massive star collapses. It fires two opposite jets moving at nearly light speed. A gamma-ray flash lasting tens of seconds is followed by an afterglow that fades from X-ray to optical to radio.
2 · Why viewing angle matters
Matter moving near light speed beams its light forward into a narrow cone, a bit like a flashlight. Look straight down the jet and it seems enormously bright. From the side it looks dim until the jet slows and the cone widens.
3 · Forward and reverse shocks
When a jet hits surrounding gas, two shock waves form. The forward shock runs outward into the gas. The reverse shock runs back into the jet's own material. Each one glows, and they emit at different wavelengths.
5/5Scientific American· Oct 30
Satellite radar survey finds sinking ground in 465 big cities
Researchers used satellite radar to measure ground sinking from 2014 to 2023 in 465 cities of over a million people. Some, including Jakarta and Mexico City, sink by up to several inches a year, mostly from groundwater pumping.
Why it matters · A sinking city is usually treated as a local engineering problem. This survey measures it the same way everywhere, which lets cities be compared and shows where nobody had looked before.
Know this first
Context from the editor, not the article · 32 sec
1 · What subsidence is
Subsidence is the slow sinking of the ground surface. In cities it is usually caused by people, not geology: pumping water out of the ground, or heavy buildings pressing down on soft soils. It tends to happen gradually, over years.
2 · Radar from orbit
Radar satellites send microwave pulses down to the ground and record the echoes. If they image the same spot again and again over time, tiny changes in how far the signal travels show the ground moving, even by a few millimeters.
3 · Why water holds land up
Water fills the spaces between grains of sediment underground and carries part of the weight above it. When that water is pumped out, the grains pack closer together and the surface drops. Clay layers often stay compacted even if the water later returns.
Hubble image of a 3-light-year-tall pillar of gas and dust in the Carina nebula, part of the interstellar medium.1 / 3
A new 3D map of the gas and dust around the sun suggests that a textbook model of the space between stars is incomplete. Jonathan Shelest, an undergraduate at the Technion in Israel, and his adviser, Shmuel Bialy, found that 41% of the nearby 1 (ISM), the thin gas and dust between stars, sits at temperatures that should be short-lived. Their paper, posted in July 2026, is still undergoing peer review.
The old picture
In 1969, three astronomers proposed that ISM gas settles into one of two stable states. Cold clouds sit near 100 2. Warm diffuse gas sits near 8,000 kelvins, kept hot by ultraviolet light from stars. Gas in between was expected to quickly cool or heat into one of the two, so very little of it should exist at any moment.
Testing this was hard. Cold hydrogen gives off no visible light. Its faint radio signal at a wavelength of 3 comes from every cloud along a line of sight at once, so the signals blend together.
How the map was built
The European Space Agency's 4 spacecraft (2013–2025) measured distances to almost 2 billion stars using 5, the apparent shift of a nearby object when you view it from two positions. Once a star's distance is known, the amount its light has dimmed and reddened shows how much dust lies in front of it. In 2024, another team used this to map dust density in 3D across a sphere more than 4,000 light-years wide around the sun.
Dust density alone doesn't give temperature. Bialy found the missing piece when he met Stanford astrophysicist Troy Porter at a conference. Porter had an unpublished 3D map of the ultraviolet light that heats the gas. The Technion team combined the two maps, about 200 gigabytes of data. At each point on a 3D grid, they compared dust density with radiation strength to decide whether the gas there is cold, warm or in between.
What they found
The split was 27% cold, 32% warm and 41% unstable. The unstable gas wasn't scattered like noise. It formed large, continuous layers wrapped around the cold clouds.
Two critics see ways the method could inflate that number. Chris McKee says cold and warm regions too close for the map to separate could blur together. Philipp Frank, also of Stanford, notes that the team's averaging method smooths sharp edges into gradual slopes, which would add in-between gas. The authors reran the analysis on a different, lower-resolution dust map and found the unstable share changed by only about 5%.
Why it matters
Galaxies turn surprisingly little of their gas into stars. If heating from stars keeps the ISM stirred up, less gas stays cold and dense enough to collapse into new stars. That could explain why models that assume calm, balanced gas don't match what astronomers observe.
A team at the Technion in Haifa, Israel, has built a 3D map of the gas and dust between stars near the sun. The map suggests that much of this material is in a state the textbooks say should hardly exist. Undergraduate Jonathan Shelest and his adviser, Shmuel Bialy, found that 27% of the local 1 is cold, 32% is warm, and 41% falls in an "unstable" middle range. They posted the paper in July 2026. Quanta Magazine reports that it is still undergoing peer review, so for now this is a preliminary result.
The quiet backdrop
The interstellar medium (ISM) is the thin gas and dust that fills the space between stars. In 2020, NASA's New Horizons probe was past Pluto's orbit when it measured a proxy for the hydrogen around it. The result worked out to about 130 hydrogen atoms per liter. Most of that gas is left over from the Big Bang. Stars add more: young stars throw out jets as they form, and dying stars shed their outer layers. Because the ISM is so thin and changes so slowly, astronomers spent about 50 years treating it as a fixed backdrop.
The standard theory dates to 1969. That year, George Field, Donald Goldsmith and Harm Habing proposed that the ISM has two stable states:
Cold clouds form where gravity pulls gas and dust together. Collisions between molecules drain away energy, so these clouds settle near 100 2 (about −173 °C).
Warm diffuse gas is too hot to collapse into clouds. Ultraviolet light from nearby stars keeps heating it, so it stays near 8,000 kelvins.
Gas at temperatures in between was expected to move to one state or the other almost at once. The reason, from general physics rather than the article, is a balance between heating and cooling. At the extremes, that balance pushes a small disturbance back to where it started. In the middle range, it amplifies the disturbance until the gas runs away to cold or warm. In 1977, Chris McKee and Jeremiah Ostriker added a third phase: plasma at about half a million kelvins, heated by supernova shock waves. Even in their version, the ordinary atomic gas outside those blast zones was supposed to be calm and in pressure balance.
Why it was hard to test
Cold hydrogen gives off no visible light. Astronomers detect it through a rare flip in the hydrogen atom, when the electron and proton switch between two states. That flip emits a faint radio wave 3 long. For more than 70 years, radio dishes have picked up this signal. But each line of sight from Earth crosses many clouds, and their signals blend together. The article compares this to looking through a stack of tinted, see-through plastic sheets: the colors merge, so you can't pick out any single layer.
How 4 made the dust map
The European Space Agency's Gaia spacecraft (2013–2025) changed what was possible. Gaia measured the distances to almost 2 billion stars using 5. Hold up your thumb and close one eye, then the other: your thumb seems to jump against the background, and it jumps more the closer it is. Gaia used the same effect, viewing stars from opposite sides of Earth's orbit around the sun. Nearby stars appear to shift more than distant ones.
Once astronomers knew each star's distance, they could measure how much its light had dimmed and reddened on the way to Earth. Dust in the ISM causes that dimming. Compare a star at 500 light-years with one at 600 light-years in the same direction: the extra dimming tells you how much dust lies between them. Do this for millions of stars and you can rebuild the dust in three dimensions.
The math is hard. Gregory Green of Westlake University in Hangzhou, China, says many different dust arrangements can fit the same data. More data helps narrow the options but makes the calculation harder. The task is to find the most physically plausible arrangement. Green and others spent a decade developing statistical methods for this. In 2024, a separate team used them to produce a continuous 3D dust-density map covering a sphere more than 4,000 light-years across, centered on the sun.
Adding heat
A dust map shows where material is, not how hot it is. Bialy got the missing piece by chance: at a conference, he met Troy Porter of Stanford University, who had made a 3D map of the ultraviolet radiation that heats the ISM but hadn't published it. The Technion team combined the two maps, about 200 gigabytes of data. At every point on a 3D grid, they compared dust density with the strength of the radiation to work out whether the gas there should be cold, warm or in between. The article doesn't give the exact conversion.
The result was the 27/32/41 split. "I thought I was wrong," Shelest said. But he found no sign of artifacts or random noise. The unstable gas formed large, continuous layers that wrapped completely around the cold clouds, which is not the scattered pattern noise would produce. Adam Ginsburg of the University of Florida reads this as a sign that something keeps stirring the gas and pushing it between states.
The caveats
Outside researchers point to two ways the method could inflate the unstable share:
Resolution. McKee says that if cold and warm regions sit closer together than the map can resolve, they could blur into what looks like in-between gas.
Smoothing. Philipp Frank of Stanford's Kavli Institute for Particle Astrophysics and Cosmology says the team used an "6," a type of average that tends to smooth over sharp contrasts. A sudden edge between a dense cold cloud and thin warm gas would come out as a gradual slope, and the gas on that slope would count as unstable.
The authors agree both problems exist, but they argue blurring can't explain most of the signal. When they ran the same analysis on a different, lower-resolution dust map, the unstable fraction changed by only about 5%. That test helps their case, but it doesn't settle it. If resolution were the main problem, a coarser map would be expected to raise the unstable share. Only a sharper map, or independent temperature measurements, can show whether the share falls as finer detail comes into view.
Why it would matter
Galaxies convert only a small share of their gas into stars, and simulations have long had trouble explaining why. If much of the ISM is constantly heated, stirred and pushed between states, less of it sits in the cold, dense form that collapses into stars. Anna Rosen of San Diego State University says 7 from stars heats and pushes on the gas around them. In her view, the classic equilibrium models give "physically grounded snapshots," but "in between those snapshots, there's still a lot going on."
The open questions are what does the stirring, how long the in-between gas lasts, and whether a sharper map will confirm the 41% figure.
Remember
The 41% figure is preliminary. The paper is unreviewed, and critics say resolution limits and averaging could inflate it.
The temperature map comes from combining Gaia's dust measurements with an unpublished 3D map of ultraviolet light.
The unstable gas forms continuous layers around cold clouds, which argues against it being random noise.
A new wiring map shows how a fruit fly's taste-sensing neurons connect to its brain and on to the neurons that control what it does next: walking, swallowing, or releasing insulin to digest food. The study, led by Carlos Ribeiro at the Champalimaud Centre for the Unknown in Lisbon, was published in Cell.
Why taste is a good place to start
Ribeiro argues that taste is a simpler problem than vision or hearing. A fly's response to a given image can vary a lot, but flies almost always go for sweet things and avoid bitter ones. In his words, "how the behavior is driven is inbuilt in the neurons." That makes it easier to follow a path from sensation to action.
How the map was made
The team used light microscopy to find every taste neuron on the tongue, throat, legs and wings of male and female flies. They worked out which taste each neuron detects from its shape and the genes it expresses. Then they traced those neurons' 1, the long output fibers, through two existing 2, complete neuron-by-neuron wiring maps: one of the whole male nervous system, published in the same issue of Cell, and one of the female brain.
What they found
Same response, different context. Sweet-sensing neurons anywhere on the body feed into similar downstream neurons, but the circuits keep track of which body part did the sensing. Ribeiro's example: sugar touched by a leg should make the fly stop, while sugar on the tongue should be swallowed.
More neuron types than the genes suggest. The 3 (throat) holds a "huge family" of taste neurons that seem to control specific actions such as swallowing and insulin release. The tongue has at least two newly identified types.
Caveats
Jan Pielage of RPTU Kaiserslautern, who was not involved, notes that the map still has to be tested. AI predicted which neurons excite or inhibit others, and the routes through several synapses were estimated mathematically, which may miss pathways made of many weak connections. Juan Sánchez-Alcañiz of Spain's research council compares a connectome to a road map: it can't tell you how traffic will flow.
The value, Pielage says, is that fly geneticists can now switch any mapped neuron on or off and watch how behavior changes.
A fruit fly tastes with much of its body. Taste-sensing neurons sit on its tongue, in its throat, and on its legs and wings, and between them they detect sweetness, bitterness, salt, water, carbonation, lipids (fats) and amino acids. A study published in Cell now maps how all of those sensors connect to the brain and back out to the neurons that drive responses: walking, swallowing, and releasing insulin to handle a meal. Carlos Ribeiro, who leads the Behavior and Metabolism Lab at the Champalimaud Centre for the Unknown in Lisbon, led the work.
"We think we know a lot about taste, but we know so little," says Jan Pielage, a neurobiologist at RPTU University Kaiserslautern who was not involved. With this map, he says, researchers can use the genetic tools available in 1, the fruit fly genus, "to really test individual hypotheses."
Why taste is a tractable problem
Following a signal from a sense organ to a behavior is hard in most systems, because the same input can lead to very different outputs. Ribeiro argues taste is an exception. The way a fly responds to a visual scene varies a lot, but flies reliably approach sweet things and avoid bitter ones. Taste neurons are specialized, so "how the behavior is driven is inbuilt in the neurons." If the behavior is largely fixed by which sensor fires, then tracing the wiring from that sensor should explain a lot about the behavior.
How the map was built
The work had two stages.
First, the team used light microscopy to find every taste-sensing neuron in the tongue, throat (2), legs and wings of both male and female flies. They assigned each neuron a taste category based on its shape and the genes it expresses.
Second, they followed each neuron's axon, the long fiber that carries its output, into the central nervous system. For this they used two existing 3, which are complete maps of every neuron and synapse built from electron-microscope images. One covers the entire male nervous system and was published in the same issue of Cell. The other is the earlier connectome of the female fly brain. Knowing where each taste axon enters these maps let the team trace the downstream neurons it reaches, and from those, the motor and hormone-releasing neurons at the far end.
Finding 1: one response, tagged by location
Sweet-sensing neurons converge on similar downstream neurons wherever they sit on the body. That fits the idea that sweetness means "accept." But the pathways also preserve which body part detected the sugar.
Ribeiro calls this "a very cool architecture," because the fly can make the same kind of decision and still adjust what it physically does. His example: "if [the food] is touched by the leg, the fly should stop, but if it's touched by the tongue, it should then be swallowed." One way to picture it (an analogy, not the authors'): every sweet signal is stamped "good food," but each stamp also carries a return address, so the brain knows whether to tell the legs to halt or the throat to swallow.
Finding 2: more diversity than the genes suggest
The pharynx contains what Ribeiro calls a "huge family" of taste-neuron types that seem to control "very specific aspects of behavior," including swallowing and insulin release. The tongue has at least two newly identified types of taste neuron.
More broadly, Ribeiro says, taste neurons are more varied than their gene expression implies: "there might be many neurons which all detect sugar, so they are molecularly similar, but they have different roles to play." Their wiring, not just their receptors, sets what they do.
Finding 3: taste lifts a brake
Taste was generally thought to be 4, meaning signals flow one way from sensor to action, and a taste would directly switch a feeding behavior on. The map suggests otherwise. "Feeding neurons are inhibited constantly," Ribeiro says, "and then what taste actually does is it releases this brake." He suggests this design may be a way to prevent runaway behavior and keep feeding under control. In practice, sugar wouldn't press an accelerator for feeding. It would take a foot off a brake that is always on. This is the team's reading of the wiring, and it still needs testing in live flies.
What the map can't yet tell you
Pielage stresses that, like any connectome, this one has to be tested and validated. Two methodological choices matter:
Predicted, not measured, signs. Connectomes show where synapses are, not whether they excite or inhibit. The team used AI to predict which 5 each neuron releases, and from that whether it excites or inhibits its targets. The brake finding depends on these predictions.
Estimated multi-step routes. For neurons linked through several intermediate synapses, the team estimated connection strength mathematically. Pielage says this "potentially underrepresents pathways built from numerous but weak connections," so circuits made of many small inputs could be missing or understated.
Juan Sánchez-Alcañiz, a tenured researcher at CSIC, Spain's National Research Council, who was not involved, offers a wider caution. Just as sequencing a genome didn't solve biology, a connectome is "a really valuable tool" but doesn't solve the brain. "It's like a road map," he says, and a map can't tell you how traffic will flow during a jam. He sees the field moving toward adding 6, which reads the genes active in individual cells, to learn which genes neurons express when they are active or inhibited, in order to "try to decipher how a nervous system is actually working."
Why it's still worth it
What the connectome adds, Pielage says, is a list of specific neurons sitting in known behavioral circuits, each of which can be tested. "In the fly, we can now activate or silence any neuron if we have genetic access [and] then see how this changes the behavior," he says. "It's basically a platform now upon which we can play." Experiments like these are how the AI-predicted signs and the brake idea will be confirmed or overturned.
Ribeiro says the roughly two-decade effort behind fly connectomics was worth it for this reason: "This lays, really, the ground for many, many future discoveries for us, and many other people." The article doesn't say which of the circuits, if any, have already been tested by switching neurons on or off.
Remember
Sweet signals from legs and tongue reach shared circuits but keep a body-part tag, so the fly stops or swallows accordingly.
Feeding neurons appear constantly inhibited; taste works by releasing that brake rather than directly triggering action.
Excitatory and inhibitory roles were AI-predicted, so the map's key claims still need testing in live flies.
Teghaza 001, an intact igneous rock from Mars's ancient crust, found in the Sahara in Mali.
A meteorite found in Mali's Sahara in 2022 has turned out to be a rare intact piece of early Mars. A team led by Yang Liu of NASA's Jet Propulsion Laboratory reports in the peer-reviewed journal Science that the rock, Teghaza 001, formed at least 4.1 billion years ago.
Why old Mars rocks are scarce
Crater counts show that most of Mars's surface is more than 3 billion years old. Yet the three main classes of Martian meteorites, 1, nakhlites and chassignites, are all 2.4 billion years old or younger. Only two ancient ones were known before: Allan Hills 84001 from Antarctica, and Black Beauty, a jumble of soil and rock fragments. Liu suggests the impacts that launch rocks off Mars act as a filter: loose or weak material doesn't survive the trip. Teghaza 001 is a single 2 rock, meaning it solidified from molten rock.
It was first labelled a shergottite. Then Carl Agee at the University of New Mexico noticed oddities, including a mineral altered on the inside but pristine outside, the reverse of weathering on Earth. It is mostly 3, a pale crustal mineral, with darker 4. Its millimetre-sized crystals imply slow cooling, about 0.01 °C per year, so the magma hardened underground rather than erupting.
What the ages reveal
5 grains, tiny crystals widely used as geological clocks, gave about 4.11 billion years. Phosphate minerals and 6, based on radioactive decay, agreed. Zircon clocks are hard to reset, and the launching impact wasn't hot enough to do it. Liu argues that hot circulating water reset them, and sodium-rich rims on the feldspar grains support this. So 4.1 billion years likely marks the rock's last soaking. It formed earlier.
Water and a deep source
The rock has about two parts per thousand more 7, a heavier oxygen isotope, than other Martian meteorites. Magma processes alone can't produce that much. Low-temperature water near the surface can. The team infers that the magma took in crust altered by surface water, or reacted with such fluids, before it finished cooling. Liu argues this fits a wet early Mars with possibly habitable environments in its crust.
Trace-element ratios differ from every other Martian meteorite. Melting models fit best if the magma's source contained 8, which is stable only more than 160 km down. That source matches none of the three to five mantle reservoirs inferred from other meteorites.
Open questions
The team suggests the rock came from the southern highlands northwest of the 9 impact basin. Orbital data can't pin a source down better than about 100 by 100 metres, though. When the rock left Mars is unknown, and other groups are now studying it. Liu says samples collected with geological context, like those from NASA's 10 rover, would tell far more.
A rock bought from Saharan meteorite hunters has turned out to be one of the oldest pieces of Mars on Earth. A team led by Yang Liu, a planetary scientist at NASA's Jet Propulsion Laboratory, reports in Science that Teghaza 001, found in Mali in 2022, formed at least 4.1 billion years ago. "It's very old and similar to some of the oldest rocks on Earth," Liu says.
A mismatch between Mars and its meteorites
Crater counts show that most of the Martian surface is more than 3 billion years old. The meteorites we have tell a different story. The three main classes, 1, nakhlites and chassignites (named after the places where the first examples fell), are all 2.4 billion years old or younger. Until now only two truly ancient Martian meteorites were known. One is Allan Hills 84001, found in Antarctica. The other is a jumbled rock from northwest Africa nicknamed Black Beauty.
Liu says nobody fully understands why old Martian rocks are so scarce on Earth, but one likely reason is how they get here. Only a large impact can throw rock off Mars, and weak material doesn't survive the trip. "If the impact hits into loose powder, it may not get cemented to become a rock and fly off Mars' gravity," she says. "So the natural impact already acts as a filter." The result is a biased collection: none of the meteorites match what orbiters see on the surface. Black Beauty comes closest, because its bulk chemistry agrees with orbital measurements. But it is a mix of soil and many small rock fragments. Teghaza 001 is a single, intact 2 rock, meaning it solidified from molten rock.
From desert stone to "very special"
Teghaza 001 arrived through the desert meteorite trade as eight stones weighing just under 800 grams in total, bought by a private collector. A piece went to the University of New Mexico, where Carl Agee and Karen Ziegler measured its oxygen isotopes. The values matched other Martian meteorites. The rock was first classified as a shergottite, the most common and youngest type. Agee noticed oddities, though. One mineral had been altered on the inside while its outside stayed pristine, the opposite of how weathering works on Earth. The oxygen isotopes looked slightly off, and there was manganese-rich material like some Liu had studied in Black Beauty. That manganese material turned out to come from Earth, but Liu says she "quickly realized the sample is very different."
The rock is mostly 3, a pale mineral common in rocky planets' crusts, with darker 4 crystals mixed in. Liu says the texture resembles some rocks in the lunar highlands. Crystals grow larger when they have more time, and these reach a millimetre across. Comparing them with lab experiments and Earth rocks, the team estimates the rock cooled by about 0.01 °C per year. That is slow enough that the magma must have hardened underground inside the crust rather than erupting.
Several clocks, one age
The team dated the rock in several independent ways. 5 grains, tiny crystals that lock in uranium and are a standard geological clock, gave about 4.11 billion years. Phosphate minerals and 6, which tracks how radioactive rubidium decays into strontium, gave similar ages. "We didn't expect that," Liu says of the agreement.
That agreement turned out to be revealing. Zircon clocks are very hard to reset. The impact that launched the rock made high-pressure minerals but wasn't hot enough to reset them. Liu argues that only a hydrothermal fluid, meaning hot water circulating through rock, could have done it. Sodium-rich rims on the feldspar grains and sodium trapped in the zircons support that idea. So 4.1 billion years probably marks the rock's last soaking, the latest of several alteration events. The rock itself crystallized earlier still.
A signature of surface water
Teghaza 001 also carries about two parts per thousand more 7, the heavier oxygen isotope, than other Martian meteorites. Neither crystallization nor gas escaping from magma can shift the ratio that much. On the other hand, water reacting with rock at low temperatures near a planet's surface produces minerals rich in heavy oxygen. The team concludes that the magma either swallowed pieces of crust that had already been altered by surface water, or reacted directly with fluids that carried the same heavy signature. Either way, water was moving through the crust and interacting with magma before the magma had finished cooling. "It's actually more dynamic and more active than we thought," Liu says. She argues this fits a wet early Mars and suggests potentially habitable environments inside its crust. That is an interpretation, not a direct detection of habitable conditions.
Clues from deep inside
The rock solidified near the surface, but its magma came from much deeper. Some trace elements are hard to move around by alteration, so their ratios preserve the source's chemistry. One of these, the ratio of zirconium to titanium, differs from that in every other Martian meteorite. The team's melting models fit the rock's chemistry best when the source contained 8, which is stable in the Martian mantle only below about 160 kilometres. Other meteorites suggest that when Mars's early global magma ocean solidified about 4.5 billion years ago, the mantle separated into three to five distinct reservoirs. Teghaza 001's source matches none of them. Liu concludes that the early mantle evolved "in a way that's more complicated than we thought."
What's still unknown
Comparing the rock's chemistry with orbital maps, the team suggests it most likely came from the southern highlands, northwest of the giant 9 impact basin. That is a best guess. At their finest, orbital instruments resolve patches of about 100 by 100 metres, often a single pixel. When the rock left Mars is also unknown. Liu says the team did "the minimum required" to show the sample is special, and two or three other groups are already studying it.
The underlying limit is that meteorites arrive with no geological context. Liu says the bigger advance would come from samples collected with known locations, like those NASA's 10 rover has gathered in and around Jezero Crater. She would especially like samples from terrain as old as Teghaza 001 or older.
Remember
Matching ages from several dating methods pointed to hot water resetting the rock's clocks, not to the rock's original formation.
Excess heavy oxygen implies surface-altered water or crust interacted with magma on Mars over 4.1 billion years ago.
Impacts filter which Martian rocks reach Earth, so the meteorite collection poorly represents Mars's mostly ancient surface.
Illustration of a gamma-ray burst's two opposite jets, which fade to longer wavelengths as they slow.1 / 4
An international team led by Yu-Han Yang of the University of Rome Tor Vergata has explained a 1 that broke the usual pattern. The study is published in The Astrophysical Journal Letters. It argues that the burst only looked weak because we saw its jet from the side.
An oddball burst
NASA's Fermi Gamma-ray Space Telescope detected GRB 260310A on March 10, 2026. A gamma-ray burst (GRB) is the flash released when a massive star's core collapses and fires two jets moving at nearly light speed. Normally a strong gamma-ray flash is followed by an 2 that fades steadily, first in X-rays, then visible light, then radio waves.
This burst did things differently. Its gamma-ray flash lasted almost a minute but carried relatively little energy. The visible-light afterglow then brightened for several hours, until it outshone the host galaxy. The X-ray and optical afterglows ranked in the top 1% of all GRBs for brightness, and the radio afterglow in the top 5%. About three weeks in, the X-rays brightened again for roughly a week before fading.
The burst was fairly close, at a 3 of about 0.153 (roughly two billion light-years). Observations from Spain's Calar Alto Observatory tied it to a 4, a star that had lost its outer hydrogen and helium layers before it exploded.
How the team pieced it together
After Fermi's alert, X-ray, optical and radio telescopes followed up, including 5, XMM-Newton, the Einstein Probe, the Large Binocular Telescope and the Very Large Array. Yang at first suspected the X-ray comeback came from another source nearby. Chandra has the sharpest X-ray vision of these telescopes, and its images ruled that out.
The team's explanation has two parts:
Off-axis viewing. Jet light is beamed forward, so a jet seen from the side looks faint at first. As the jet slows, its light spreads over a wider angle and the afterglow seems to brighten. That explains the weak flash and the slow rise.
A two-component jet. The jet has a narrow, bright core inside a wider, dimmer sheath. The radio comes mostly from a 6, which runs back into the jet's own material. The X-rays come mostly from a 7, which runs out into the surrounding gas. The standard single-jet models with these two shocks could not fit the data.
Caveats
The authors call the late X-ray brightening a "critical test" of their model. They say it might instead come from fresh energy being pumped into the shock. If it really is the narrow core coming into view, the core would have to be extremely powerful. Yang estimates the jet carried about a billion years of the Sun's energy output into a pencil-thin beam. Seen head-on, it might have looked like a monster explosion.
An international team led by Yu-Han Yang, a postdoctoral scholar at the University of Rome Tor Vergata, has published an explanation for one of 2026's odder cosmic explosions. The paper is titled "An Extremely Bright Slow-rising Afterglow from an Off-axis Jet in GRB 260310A" and appears in the peer-reviewed Astrophysical Journal Letters. Its claim is that a burst which looked weak at first was a very powerful jet that we happened to see from the side.
What 1 normally do
A gamma-ray burst (GRB) is among the most energetic events known. Long GRBs, the kind studied here, come from the collapse of a massive star's core, which leaves behind a neutron star or black hole. The collapse launches two opposite jets of plasma moving at nearly the speed of light. A separate class, short GRBs, has a different origin.
A typical long GRB starts with a gamma-ray flash of about 20 to 30 seconds, called the prompt emission. An 2 follows as the jet ploughs into the surrounding gas and slows down. It shows up first in X-rays, then in visible light, then in radio waves, and generally gets fainter over time. Satellites spot hundreds of GRBs a year. Most are seen with a jet pointed nearly straight at Earth, because that is when they look brightest.
Why GRB 260310A stood out
NASA's Fermi Gamma-ray Space Telescope caught GRB 260310A on March 10, 2026. The name encodes the date. The burst broke the usual pattern in four ways:
Weak prompt emission. The gamma-ray flash lasted almost a minute but released relatively little energy. The paper also notes a "hard peak energy": the gamma rays peaked at high photon energies. Faint bursts usually peak at lower energies, so this combination was odd.
A slow-rising optical afterglow. Visible light from the afterglow brightened gradually for several hours instead of fading, until it outshone the host galaxy.
Extreme afterglow brightness. The X-ray and optical afterglows ranked in the top 1% of known GRBs, and the radio afterglow in the top 5%. The authors call it "one of the brightest afterglows ever observed in the X-ray, optical, and radio bands."
A late X-ray comeback. After fading quickly, the X-rays brightened again at around 20 to 30 days, stayed bright for about a week, and then faded slowly. They were still detectable months later.
The burst sits at a 3 of about 0.153, roughly two billion light-years away, which is close by GRB standards. Observations from Calar Alto Observatory in Spain linked it to a 4. That is a star stripped of its outer hydrogen and helium layers, whose debris expands very fast. Such supernovae are the usual partners of long GRBs.
Co-author Roberto Ricci led the radio observations. "We could explain individual features of this burst, but fitting them all into the same picture proved much harder," he said.
Ruling out a mix-up
The team followed the burst across many wavelengths:
X-ray:5, XMM-Newton and the China/ESA Einstein Probe.
Optical: the Large Binocular Telescope.
Radio: the Karl G. Jansky Very Large Array.
Yang coordinated the X-ray monitoring. He first suspected the late X-ray rise came from an unrelated source nearby, blended into the same blurry X-ray image. Only Chandra has fine enough angular resolution to separate close sources. The team used some of its Director's Discretionary Time, observing time the observatory director can award for urgent targets. Chandra showed no contaminating source, so the rebrightening belonged to the afterglow.
Mechanism, part one: looking from the side
The key physics is 6. Matter moving close to light speed concentrates its light into a narrow cone along its direction of travel. Yang uses a flashlight analogy: "the jet appears brightest when seen head-on." An observer outside that cone sees much less, which explains the weak gamma-ray flash.
As the jet hits surrounding gas and slows, its beaming cone widens. Eventually the cone sweeps across an observer who is off to the side. To that observer, the afterglow seems to rise and peak late. That matches the slow optical rise. Yang says that as the jet slowed, "its light became visible over a wider range of angles, allowing us to probe the jet's shape, from its outer edges toward its inner core." The authors write that weak prompt emission, hard peak energy and late afterglow onset together are "naturally explained by a GRB jet viewed off-axis."
Mechanism, part two: two shocks and a two-layer jet
When a jet hits surrounding gas, two shock waves form:
The 7 runs outward into the gas.
The 8 runs backward into the jet's own material.
The team's analysis of the broadband spectrum favours both. A bright reverse shock dominates the radio and reaches into the optical. A forward shock contributes to the optical and dominates the X-rays.
Even so, "none of our standard RS plus FS models" fit the data, the authors write. The late X-ray rise in particular is "difficult to accommodate within a single FS model." They offer two options:
A late burst of energy injected into the shock.
Extra structure in the jet itself.
They favour the second, a two-component jet: a narrow, bright inner core wrapped in a wider, slower, dimmer sheath. Here is how that plays out for a side-on observer. The wider sheath points closer to our line of sight, so its light reaches us first and produces the early peak. The narrow core points further away, so its emission only comes into view weeks later as it slows, producing the X-ray comeback. The authors argue that a slow jet seen roughly head-on "would not be able to combine the early peak and late rebrightening," while "an off-axis two-component jet can accommodate both these peaks."
What remains open
The authors are careful not to overstate this. They call the late rebrightening "a critical test for the two-component jet model." If it really is the narrow core emerging, that core must be extraordinarily powerful, implying "extreme luminosities and energetics for an on-axis observer." Yang's estimate is that the explosion packed roughly a billion years of the Sun's energy output into a pencil-thin beam. "What appeared to us as an unremarkable GRB might have looked like a monster explosion to an observer looking straight down the jet's core," he says. The energy-injection explanation has not been fully excluded.
The wider point concerns bias. "Most of our models are calibrated using events seen from that same perspective," Yang notes, meaning head-on. If GRB models are built mainly from head-on bursts, side-on events could be misclassified as intrinsically faint. A nearby, well-observed burst like this one helps measure how jets are actually structured. This is one event, though, and the article does not say how common such side-on bursts might be.
Remember
How faint a burst looks can depend on viewing angle, not on how much energy it released.
The late X-ray rebrightening suggests a narrow, powerful core inside a wider, dimmer jet sheath.
The authors say an alternative, late energy injection into the shock, is not fully ruled out.
World map of 465 large cities, colored by sinking severity score (0–100) from 2014 to 2023.1 / 2
The finding
A study in the peer-reviewed journal 1 used satellite radar to measure ground sinking in 465 urban areas with more than a million people each, from 2014 to 2023. Some, including New Orleans, Mexico City, Jakarta and Rome, are sinking by up to several inches a year. Coauthor Guoquan Wang, a University of Houston geophysicist, calls it "a global urban problem," not a local curiosity.
How they measured it
2, the slow sinking of land, is hard to track from the ground, where instruments measure single points. Instead, the team compared radar images of the same areas taken by satellites over many passes. If the ground drops even a few millimeters, the radar echo takes slightly longer to come back. The method is generally called 3 (interferometric synthetic aperture radar). The researchers looked at patches only a few houses wide. Roland Bürgmann of UC Berkeley, who wasn't involved, calls it the most comprehensive survey of urban subsidence so far.
Where and why
Sinking was worst in Southeast Asia, West and Central Asia, and North America. It also turned up in South America and Africa, two regions with little previous research. For almost half the cities, no satellite radar study had ever been done.
The main cause is usually groundwater pumping by utilities and industry. Underground water helps hold the soil up, much like air in a cushion (the article's analogy), so removing it lets the ground sag. In some places, such as New York City, the weight of buildings compressing soft soil adds to the problem.
Uneven sinking does the damage
Bürgmann says a whole city sinking evenly by an inch wouldn't matter much. The trouble comes when neighborhoods sink at different rates. The strain cracks foundations, breaks sewer lines, damages subways and worsens flooding. Jakarta is so badly affected that Indonesia is building a new capital on Borneo.
It can be stopped
Wang says the sinking is "cumulative and largely irreversible," but cities can halt it. Tokyo once sank faster than many cities do today. It stopped after officials regulated groundwater pumping and switched to 4. Beijing and Tianjin replaced groundwater with water piped from the Yangtze River basin, and sinking has halted in parts of that region. "In most places, we have plenty of time to react," says Manoochehr Shirzaei, a Virginia Tech scientist who peer-reviewed the study, "but we have to do it."
The article gives few per-city figures and doesn't explain how the severity rankings were calculated.
A global measurement, not a local one
Plenty of individual cities are known to be sinking. A new study in the peer-reviewed journal 1 is the first to measure the problem the same way across the whole urban world. Researchers used satellite radar to track 465 urban areas, each with more than a million people, from 2014 to 2023. Hotspots include New Orleans, Mexico City, Jakarta and Rome. In the fastest-sinking places the ground drops by up to several inches a year. The results are also published as an interactive map.
Coauthor Guoquan Wang, a geophysicist at the University of Houston, says sinking turns out to be "a global urban problem," not "a local curiosity" confined to a few cities. Roland Bürgmann, a geophysicist at UC Berkeley who was not involved, calls it "the most comprehensive and systematic exploration of subsidence in urban cities."
How radar measures millimeters from orbit
2, the slow sinking of land, is hard to monitor from the ground. Surveyors and GPS stations measure single points. A city needs coverage of every block, repeated for years.
The study uses satellite radar instead. A satellite takes radar images of the same area on repeated passes and compares how long the signal takes to travel. If the ground has sunk even a few millimeters (a millimeter is about a twenty-fifth of an inch), the echo comes back slightly later than before. The technique is generally called 3, short for interferometric synthetic aperture radar. Strictly, it compares the phase of the returning waves, which is how it detects distance changes far smaller than the radar's own wavelength. Stack many passes over a decade and you get a sinking rate for each patch of ground.
The team worked at fine resolution, looking at chunks of land only a few houses wide. That detail matters because, as the next section explains, the variation between neighborhoods is often what does the damage.
The article doesn't name the satellites. The 2014 start date matches the launch of the European Space Agency's 4 radar mission, whose free, repeated global coverage makes surveys like this feasible. That link is an inference, not something the article states.
Where the ground is dropping
Sinking was most severe in Southeast Asia, West and Central Asia, and North America. It also showed up in South America and Africa, two regions with little previous subsidence research. For almost half of the 465 cities, no satellite radar study had ever been done. That may be the survey's most useful result: it fills in a map that was largely blank.
The figures summarize each city in two ways. One is a 5 from 0 to 100. The other compares the median sinking rate with the rate of the fastest-sinking 5 percent of the city's land. The gap between those two numbers shows how unevenly a city is sinking.
Why it happens
The causes vary. In New York City, the weight of buildings compresses the soft soils underneath. The biggest driver almost everywhere, though, is groundwater pumping by water utilities, industry and others. The article compares groundwater to the air in a cushion: it helps hold up the soil, and when you pump it out, the ground slowly sags.
Why uneven sinking is the real danger
Bürgmann points out that if a whole city dropped smoothly by an inch or so, "it wouldn't be that big of a problem." The trouble starts when one neighborhood sinks faster than the next one. The difference bends and stretches everything that spans the boundary. Over time that strain cracks foundations, breaks sewer lines, damages subway systems and makes flooding worse.
Jakarta is the extreme case. Sinking and flooding there have become so severe that Indonesia is building a new capital city on the island of Borneo.
It can be stopped, but not undone
Wang calls subsidence "cumulative and largely irreversible." The usual explanation is that compacted clay layers don't spring back when water returns. But cities have stopped it from getting worse:
Tokyo once sank "way faster than many cities today," says Manoochehr Shirzaei, a Virginia Tech geoscientist who peer-reviewed the study. It stopped after officials regulated groundwater pumping and switched to 6 sources.
Beijing and Tianjin stopped pumping groundwater and replaced it with surface water piped from the Yangtze River basin. Wang says sinking has since halted in parts of the region.
"In most places, we have plenty of time to react," Shirzaei says, "but we have to do it."
What the article leaves open
The coverage is brief. It doesn't give per-city numbers beyond the hotspot names. It doesn't explain how the severity score is built, how the researchers handled satellite data gaps, or how they separated pumping from other causes in each city. It also doesn't say whether the 2014–2023 trends are speeding up or slowing down. And the fixes it describes, regulating pumping and importing surface water, depend on a city having another water source. For many of the newly surveyed cities in Africa and South America, the article doesn't say whether one exists.
Remember
Almost half of the 465 cities had never had their sinking measured by satellite radar before.
Differences in sinking between neighborhoods do more damage than the sinking itself.
Tokyo, Beijing and Tianjin show that switching from groundwater to surface water can halt subsidence.
Taste releases a brake. Taste was thought to work purely 4, with signals flowing one way from sense to action. Instead, the feeding neurons are held back constantly, and taste lifts that inhibition. Ribeiro suggests this may stop behavior from running out of control.
