San Diego County, California. Five thousand five hundred feet up, on a forested ridge in the Cleveland National Forest, sits the new 200-inch Hale telescope — the largest ever built. U.S. Geological Survey, Santa Ana sheet, 1:250,000 — surveyed 1949.
Palomar Mountain, California. November 1949.
Five thousand feet above the orange groves and the Navy airfields of San Diego County, the most ambitious map of the sky ever attempted is about to begin.
George Abell is a graduate student. This is his first job in astronomy — observer on the sky survey. From these plates, he will discover 2,712 galaxy clusters. But tonight, he just needs clear skies.
Each field is photographed twice. First a red-sensitive plate — fifty minutes. Then a blue plate — twelve minutes. Two portraits of the same stars, in two colors of light, taken minutes apart.
For each exposure, someone must sit at the guide telescope and keep a star centered by hand. Fifty minutes. In the cold. Night after night, for seven years.
After exposure, the plate rides a dumbwaiter to the darkroom. Development in absolute blackness. No safelights — the emulsion sees every color of light.
1,620 exposures will be taken. Only 935 pairs will pass inspection. The survivors go into drawers. They will sleep there for decades.
An anomaly on plate XE 325, and its blue companion. Can you see it? No one detects it at the time. The plates go into the drawer.
THE ATOMIC FLASH — 5:15 AM
Palomar Mountain at dawn. The dome is closed, the night's work done. On the northeast horizon — a sudden bloom of white-orange light. An atomic flash from the Nevada Test Site, 275 miles away. For a moment, the fir trees cast sharp shadows. Then darkness returns. The sky above is still full of stars.
And in the desert to the northeast, the United States is splitting atoms. The flash from Yucca Flat is visible from the mountain. The astronomers barely notice. They've been watching the sky all night — but looking the other way.
The plates sleep in their drawers
70 years later. 2017.
Chapter 1
The Vanishing
Uppsala · Zurich · The Canary Islands
Dr. Beatriz Villarroel. A Swedish physicist turned astronomer. From her laboratory at NORDITA in Stockholm, she studies the violent hearts of galaxies — quasars, black holes, things that burn.
If a massive star collapses directly into a black hole — no supernova, no explosion — it would simply vanish from the record. Has anyone actually looked?
The loud death.
Most massive stars die loud — a supernova so bright it can be seen across galaxies. Light. Heat. A funeral with fireworks.
The quiet death.
But what if one died quiet? Collapsed into a black hole without a flash? It would simply — vanish.
She proposes a research project: cross-match the sky as it was photographed in the 1950s with the sky as we see it today — and find the stars that have disappeared into black holes. She calls it VASCO. Vanishing and Appearing Sources during a Century of Observations.
600 million objects.
She writes software to compare each speck of light in the 1950s sky with the sky we see today — looking for the ones that aren't there anymore.
Tenerife. Feb 2020.
On her screen: Plate XE 325 — exposed at Palomar on April 12, 1950. The plate that went into the drawer.
Villarroel
I was sitting there with my office mate in Spain... and I was just wondering, "So what is it? What are we seeing?"
— Penn State, "My Personal Journey Through the Unknown"
She sees them. Nine pinpoints of light. Clustered together on a single 1950s plate.
Exhibit A — Plate XE 325
Nine sources, clustered in a single patch of sky. Appearing within half an hour on a fifty-minute red exposure. Absent on the blue companion plate taken half an hour earlier. Absent on the next red plate of the same field, six days later. Absent in every modern survey. Gone forever.
Villarroel
What are we looking at? Nine sources of light, all there within half an hour — and then gone. What are they?
Nine point sources. Nine moments of light, photographed on a single mountain on a single night. Here they are, one by one. Real plate cutouts.
TRANSIENT 1
TRANSIENTS 2 & 3
TRANSIENTS 4, 5, 6
TRANSIENTS 7 & 8
TRANSIENT 9
Each one observed for fifty minutes. Each one gone in the next exposure. None of them visible in the modern sky. — Cutouts from Villarroel et al. (2021), Scientific Reports, CC BY 4.0
"No satellites are known to have existed prior to the Soviet-made Sputnik in 1957 — seven years after the appearance of the transients in the 1950 POSS-I image."
"…more likely explained by a Solar system satellite of artificial or natural origin."
"We believe the mystery of the simultaneous transients is a detective story worth of the attention of the astronomical community."
— Villarroel et al. (2021), Scientific Reports, p. 6. Highlights added.
The paper has done what good science does: it has given the world something it cannot easily explain — and forced everyone, supporters and skeptics alike, to look harder.
After the paper, the search opens up.
She recruits thirty volunteers from six countries — not astronomers, just careful eyes — and teaches them to look. Together they examine fifteen thousand image pairs. The nine on Plate XE 325 weren't a fluke. The catalog grows.
QUIET MOMENT — NORDITA at night
Villarroel alone in her office at NORDITA. Night. Stockholm city lights through the window. She's looking at the screen but not really seeing it. Thinking.
Villarroel
Even if this turns out to be some new physical phenomenon, that's super exciting. That would mean we have discovered something new that nobody knew existed.
"It only takes one to be real and it changes humanity for ever."
— The Guardian
A bold claim is not enough. The science has to hold.
Chapter 2
The Shadow Test
First, the critics.
Edinburgh.
Dr. Nigel Hambly. Royal Observatory Edinburgh. He has a hypothesis — and he can prove it without ever looking at the sky.
Frame 1 · The Scanner
SuperCOSMOS, Edinburgh. The machine that turned the 1950s sky into pixels. Almost everyone who searches the POSS-I archive is using its output.
Frame 2 · The Copies
But Edinburgh never scanned the originals. They scanned glass duplicates. The originals still sleep in their drawers at Palomar.
Hambly's argument: some of these "transients" may not be sky at all. They are dust, emulsion holes, fibres — introduced when the originals were copied to glass, or scratched in by decades of handling. Picked up by the scanner. Catalogued as stars.
"There'd be nobody happier than me if they are right. But I suspect they are wrong."
— Dr. Nigel Hambly, Royal Observatory Edinburgh
Tenerife. The week the paper landed.
Villarroel
We need more data. Either we shore this up — or we put the artifact problem to bed for good.
HER ANSWER — A TEST THAT DOES NOT NEED CLEAN PLATES.
Whether the plates are pristine or filthy, geometry is geometry. By now Villarroel's catalogue has grown well beyond the original nine — over 100,000 short-lived transients picked out of the POSS-I plates. She proposes a test that contamination cannot fake.
If the transients are sunlight glinting off objects in orbit, they must obey the Sun. When Earth's shadow falls across an orbit, anything reflective inside that shadow has no light available to reflect, so it goes dark. So with a hundred thousand candidate transients in her dataset, fewer of them should land inside Earth's shadow than pure chance would predict.
Sun on the left · Earth in the middle · Satellite in orbit · Palomar watching
In plain language: out of every thousand transients, you'd expect about eleven to land inside Earth's shadow by chance. Only three actually do. More than three times fewer than the sky should contain.
The 21.9 σ figure is statistician-speak for this is not a coincidence. The chance of getting a deficit this big from random data is, for all practical purposes, zero.
The conclusion is hard to escape: these flashes are real, reflective objects orbiting the Earth.
** As it happens, this result was independently replicated a year later by Doherty et al. (2026).
The plates with the most transients seemed to be the ones taken close to the days the United States detonated atomic weapons. A hunch. She needed help to test it.
Nashville. Vanderbilt University.
Dr. Stephen Bruehl. Clinical psychologist. He studies pain — the fuzzy, noisy human kind. Villarroel sends him a drive: 2,718 nights of POSS-I transients. She wants him to test her hunch.
HE PUTS HER HUNCH UNDER A MICROSCOPE.
HER HUNCH WAS RIGHT.
On a typical Palomar night with no nearby test, transients show up on about 11% of nights.
The day after a test, that jumps to 18.5%.
Same instrument. Same plates. Same observers. The bombs are the only variable.
THE SHADOW FILTER
Then they ran it again — but this time, only on the transients that had passed Villarroel's shadow test. The ones that need the Sun. The ones that look like real, reflective objects.
"The magnitude of the association between these flashes of light and nuclear tests was surprising — as was the very specific time at which they most often occurred: namely, the day after a test."
— Dr. Stephen Bruehl, Vanderbilt University Medical Center
ONE DATE KEPT TURNING UP.
While searching the plates, Villarroel and her team kept finding transients on a particular night — including a triple flash that vanished within fifty minutes on a single plate. The plate was exposed at Palomar, 08:52 UT, July 19 1952.
“Apparently, in 1952, during two consecutive weekends — on the 19th of July and the 27th of July — there was the most famous UFO sighting probably during the last hundred years over Washington. And it was so big that even the US Air Force had to make a special press conference.”
Barnes · National Airport
We knew immediately that a very strange situation existed… their movements were completely radical compared to those of ordinary aircraft.
Airman Brady · Andrews AFB
An object which appeared to be like an orange ball of fire, trailing a tail… unlike anything I had ever seen before.
Quotes via Project Blue Book / E.J. Ruppelt, The Report on Unidentified Flying Objects (1956). Wikipedia summary →
Albert Chop · USAF Press
Lt. Patterson radioed for instructions when the objects surrounded his fighter. Nobody answered. Because we didn’t know what to tell him.
Contemporary 1952 news comic, "Saucers Over Washington, D.C."
U.S. National Archives, public domain. The Air Force did not permit photographs of the actual radar scopes that night.
Palomar Mountain, California. The same night.
2,500 miles west. Same sky. The 48-inch Schmidt camera was running a fifty-minute red exposure. The observer had no radio to Washington. He didn't know what had just been photographed.
No one noticed. No one would — for seventy years.
THE SAME NIGHT, AT 08:52 UT.
Three pinpoints of light. One photographic plate. Vanished within fifty minutes. The Palomar archive logged them. The newspapers logged what flew over the White House later that night. Same date.
Top-left: the triple flash, just above center. Top-right: 56 minutes later. Gone. Bottom row: two months later. Still gone.
Dallas, Texas. Independent researcher. Statistical analysis · data analytics.
No university, no lab, no funding. He pulled down the dataset, wrote his own code from scratch, and ran every test independently. The findings held up — both the nuclear-test correlation and the Earth-shadow deficit.
Maybe it was only Palomar. Maybe one telescope, one mountain, one drawer of plates. Busko looked at a different archive entirely — the digitised plates from the Hamburg Observatory’s Schmidt camera, mid-1950s. He found the same narrow, star-like flashes. Different telescope. Different continent. Same signature.
Independent researcher Former US Navy reactor operator (USS South Carolina)
Cann asked a different question: what happens when Earth’s magnetic field gets battered? He cross-referenced the transient dates with the geomagnetic Kp index. The flashes drop sharply during strong magnetic storms. Then, twenty-five to forty-five days later — once the field has calmed and the plasma has refilled — they surge back to roughly three times baseline.
A camera defect can’t care about Earth’s magnetic field. But something trapped in that field can.
Cann (2026), arXiv:2604.04950 → "Geomagnetic storm suppression of photographic plate transient detections in the POSS-I archive"
THE STRONGEST CHALLENGE
The signal itself may be illusory.
Watters, Dominé, Little, Pratt & Knuth (2026)
Not every astronomer is convinced the underlying detections survive proper scrutiny. A reanalysis by the Watters team argues:
The Earth-shadow deficit disappears in their reanalysis.
The linear-cluster patterns are explained by ordinary catalogue stars at the survey edge.
The nuclear-test correlation vanishes once you normalise by which dates Palomar was actually observing.
Both sides agree on the raw photographic plates. They disagree about how to count what’s on them.
Villarroel
I cannot find any other consistent explanation other than that we are looking at something artificial before Sputnik 1. For me, this looks technological. But I may be wrong.
Based on peer-reviewed research published 2020–2026.
All quotes attributed to their original sources. The Palomar Lights — a story told in data, glass, and light.
DIGGER
A comics imprint for true science stories. One issue at a time.