Today’s space digest put out one line that hides an entire detective saga: “James Webb finds the source of the most distant radio burst.” For the first time, astronomers have precisely located the host galaxy of the most distant known fast radio burst—mysterious millisecond flashes that almost never repeat.
The headline is honest, but incomplete. Behind it lie twenty-five years of history in which the chief antagonist of progress turned out to be neither a black hole nor dark matter, but two microwave ovens in the cafeteria of an Australian observatory. Ovens that fooled the planet’s best radio astronomers for seventeen years and nearly had one of the greatest discoveries of the 2000s written off as junk—before it had even properly happened.
A fast radio burst (FRB) is a pulse of radio emission lasting milliseconds, arriving from beyond our galaxy. In that time, its source manages to release energy comparable to the Sun’s output over several days. All the energy humanity uses in a year is nothing next to one such “click,” shorter than the blink of an eye.
A burst comes with a built-in ID—its dispersion measure. Radio waves of different frequencies travel through interstellar gas at different speeds: high frequencies arrive a little earlier, low ones a little later. From the difference in delay, astronomers calculate how much ionized matter the burst passed through on its way to us. For extragalactic bursts, this measure is enormous: the signal has had time to filter through billions of light-years of thin plasma. That is exactly what made FRBs both priceless probes of the cosmos and a verification nightmare: any terrestrial source of radio interference that happens to reproduce a similar frequency shift looks indistinguishable from an event at the edge of the Universe.
It all began with bureaucracy, as it often does in science. In 2007, graduate student David Narkevic at West Virginia University was combing through archival data from a pulsar survey conducted six years earlier with Australia’s Parkes radio telescope—and in a recording from July 24, 2001, he found a burst that fit no model. A pulse lasting around five milliseconds, a hundred times brighter than anything expected, with a dispersion measure pointing to a source beyond the Milky Way.
The group’s leader, Duncan Lorimer, and his colleague Maura McLaughlin took the finding to Science: the paper “A Bright Millisecond Radio Burst of Extragalactic Origin” came out in September 2007. The community’s response was predictable: telescopes had caught interference masquerading as space before—and this signal looked too much like one big data-processing mistake. Skepticism ran so deep that for six years the discovery hung by a thread: many people simply were not convinced.
Then events followed the detective-story script: a serial “suspect” appeared, one that nearly became the verdict.
As it turned out, old Parkes data were full of such signals. Sarah Burke-Spolaor, who dug through the archive after Lorimer’s publication, found sixteen bursts suspiciously similar to FRBs in spectrum and shape. One detail stood out: a real extragalactic signal appears in one or two beams of a multibeam telescope—it is point-like. These bursts, meanwhile, flooded the entire field of view. The source was nearby, on Earth. But what was it?
That is how the term “peryton” was born. Burke-Spolaor named the mysterious signals after a creature from Jorge Luis Borges’s Book of Imaginary Beings—the peryton, a winged deer said to have appeared in a now-lost medieval manuscript. The mythical beast had one unsettling trait: it cast not a deer’s shadow, but a human one. Something strange made by humans—the perfect metaphor for signals that imitated space while remaining terrestrial.
Perytons had been observed for a long time: the earliest were caught back on June 23, 1998—sixteen of them in seven minutes. Over seventeen years, Parkes counted forty-six. They came mostly during the day, almost never at night or on weekends. Hypotheses multiplied: lightning, aircraft, satellites, ionospheric tricks. Each fell apart under scrutiny. And the shadow of skepticism fell ever more heavily on the FRB phenomenon itself: if Parkes—the dish that relayed television footage of the Apollo 11 Moon landing in 1969—was catching ghosts, then who could be trusted?
The resolution came in 2015, when graduate student Emily Petroff took on the perytons. Her job was to search Parkes data for new FRBs in real time, and her supervisor set one condition: “That’s great, but have you sorted out the perytons yet?” They had not.
Petroff and her colleagues installed an additional receiver on the telescope to monitor interference in the high-frequency band, then waited. The picture came together quickly: perytons happened only during working hours—and suspiciously often around lunch. Before every burst, the monitor caught an outburst of radiation at 2.5 GHz, precisely the frequency at which microwave ovens operate. After that, it was pure engineering: one engineer stood by the oven with a radio, another by the telescope, and on March 17, 2015, after heating water in ceramic mugs and opening the door before the timer clicked off, they reproduced three perytons in a row.
The mechanism was indecently elegant. When a microwave door is opened before the cycle has finished, the magnetron, in its shutdown phase, spits out a short pulse around 2.3–2.5 GHz. That signal bleeds into Parkes’s operating band and there looks like a burst sweeping across frequency—a precise copy of the dispersion slope of a real cosmic signal. The culprits were two Matsushita microwave ovens in the observatory cafeteria, both more than twenty-seven years old. If their doors were opened early, each one “fired” a peryton roughly half the time.
The irony of the ending: by exposing the source of the interference, the team also gave FRBs an alibi. The parameters of the perytons differed so radically from the Lorimer burst that attributing it to a microwave would have meant turning physics inside out—the dispersion measure of FRB 010724 would have required an impossible path through plasma. The 2001 burst remained cosmic. And in 2013, before the oven case was even solved, Thornton’s group published four new bursts in Science with the “right” dispersion measures—the phenomenon stopped being a one-off curiosity and became a population.
The next chapter required breaking the main stereotype: an FRB does not have to be a one-off. In 2012, the Arecibo radio telescope recorded a burst, and in 2016 astronomers announced a sensation: the source repeats. The first “repeater,” FRB 121102, was reliably tied to a dwarf galaxy at redshift 0.193: bursts are born in a galaxy, not in some kind of “vacuum energy fluctuations.”
And on April 28, 2020, what happened made the theory almost ironclad: Canada’s CHIME and America’s STARE2 simultaneously caught a bright millisecond radio burst from our own galaxy. Its source was magnetar SGR 1935+2154—a neutron star with a monstrous magnetic field, the strongest known in the Universe—lying a few tens of thousands of light-years away. The burst arrived paired with a hard X-ray pulse.
This was the first FRB tied to a specific known object. Magnetars—and, apparently, their “starquakes,” rearrangements of the crust that release magnetic energy—became the prime suspects in producing the mysterious clicks. The April burst was weaker than extragalactic ones, but suspiciously similar to them in every way—as though nature had deliberately left us a scale model close to home.
And now—the ending to date. On October 8, 2026, an international team announced in Science that Webb and the South African MeerKAT radio telescope had revealed the origin of burst FRB 20240304B, caught by MeerKAT on March 4, 2024. The NIRCam camera found its host galaxy; the NIRSpec spectrograph measured its redshift: z = 2.148. The burst’s light had traveled to us for more than ten billion years—the Universe at that moment was an infant around three billion years old, in the era of “cosmic noon,” the peak of star birth.
And here is the main surprise: the host galaxy is a dwarf, roughly a thousand times less massive than typical FRB host galaxies. Its stars were born literally thirty million years before the burst. That combination—tiny mass and suddenly flaring star formation—crosses neutron-star mergers off the list of prime suspects: such a merger needs billions of years for the stars to draw together, while here everything happened instantly by astronomical standards. What remains is a magnetar born in that same brief episode of star formation: a young neutron star flared before its galaxy had time to grow old. “Our work says it is highly unlikely that this burst was produced by a neutron-star merger,” sums up team leader Manisha Caleb.
As a bonus, the burst served as a probe of matter along the way: its dispersion measure sounded out a previously unknown galaxy cluster at redshift 0.3 and the Virgo Cluster, 54 million light-years from us. Distant bursts are now being caught regularly—and every case refines the map of the Universe’s “hidden” baryonic matter. Physics that spent seventeen years learning to tell space from a microwave now uses it to measure the Universe.
What gets me in this story is exactly what Borges packed into the image of the peryton: a strangeness casting a human shadow. A twenty-seven-year-old oven in an observatory cafeteria proved capable of perfectly copying the identifying signature of an event from the other end of the Universe. It took seventeen years to separate a kitchen magnetron from magnetars. And that very painful experience of verification bore fruit: today’s news of the most distant burst became possible because, once upon a time, astronomers learned not to trust beautiful signals.
There is an ominous note here too: the microwave was caught, but it has been replaced by satellite megaconstellations—thousands of transmitters whose radio noise is a problem an order of magnitude harder than two ovens in a cafeteria. So radio astronomers’ habit of suspecting every beautiful signal will remain useful: first the microwave, now the satellites—the lesson of the perytons works day after day. Asking, “Is this definitely space, or is it someone nearby?” is not paranoia, but the very essence of the profession.