About Verant Files

Verant Files investigates cases that were never closed.

What this is

Every day we publish a short film about one unsolved case: a cosmic anomaly no instrument can explain, or a human disappearance no investigation ever resolved. The short is the opening. This site is the file behind it.

Each case file is researched from primary and reputable secondary sources, every source is linked at the bottom of the page, and where the facts are disputed we say so rather than choosing the more dramatic version.

How FILE requests work

Comment FILE on any video and that case moves up the queue. The most requested cases get the full written investigation here first, and then the long-form deep dive on the channel. The queue is genuinely ordered by what the audience asks for.

Corrections

If you find an error, or you have primary material on an open case, write to hola.verant@gmail.com. Corrections are made in place and noted.

Where to find us

YouTube · TikTok · Instagram

Planet Nine

A world several times Earth's mass may be steering the outer Solar System. Ten years on, no one has seen it.

The Hook

On 20 January 2016, the Astronomical Journal published a paper by Konstantin Batygin and Michael Brown of Caltech with a title that read like a claim of discovery: "Evidence for a Distant Giant Planet in the Solar System." No image accompanied it. There was no photograph, no dot on a plate, no radio echo. There was only a pattern.

The six most distant objects then known to orbit entirely beyond Neptune, on paths that take them hundreds of times farther from the Sun than Earth, all swing out in roughly the same direction. Seen in three dimensions, their orbits also tilt away from the plane of the planets by a similar angle. Caltech's own announcement put the odds of that happening by chance at about 0.007 percent.

Batygin and Brown said the simplest explanation was a planet around ten times Earth's mass, orbiting some twenty times farther out than Neptune, shepherding those bodies with its gravity. Brown is the man whose discoveries got Pluto demoted. Now he was proposing a replacement. Ten years later the pattern is still there, still contested, and the planet has still not appeared.

The Facts

The case rests on a small population of bodies called extreme trans-Neptunian objects, or ETNOs. They have semi-major axes greater than about 150 astronomical units (AU; one AU is the Earth-Sun distance) and perihelia, their closest approaches to the Sun, well outside Neptune's reach at 30 AU. The first and most famous is Sedna, found in 2003 by Brown, Chad Trujillo and David Rabinowitz, with a perihelion of 76 AU. The second, 2012 VP113, was announced by Trujillo and Scott Sheppard in Nature on 27 March 2014, with a perihelion of 80 AU. In that paper the pair noted that Sedna, 2012 VP113 and other distant bodies shared a clustering of one orbital angle, the argument of perihelion, and floated a super-Earth of two to fifteen Earth masses at 200 to 300 AU as a possible cause.

Batygin and Brown took that observation further. In their 2016 analysis the six objects with the largest orbits were clustered not only in argument of perihelion but in the physical orientation of their long axes and in their orbital tilt. Their simulations showed that a massive planet on an eccentric orbit anti-aligned with the cluster, pointing the opposite way, could hold the configuration stable over the age of the Solar System. The same planet, they argued, would naturally produce other oddities: objects on orbits tilted almost perpendicular to the planets, and detached bodies such as Sedna that Neptune cannot explain.

The hypothesis has been refined since. In "The Orbit of Planet Nine," published in the Astronomical Journal in 2021, Brown and Batygin used a larger sample and a treatment of survey biases to derive a best estimate: a planet of 6.2 Earth masses (plus 2.2, minus 1.3), on an orbit with a semi-major axis of 380 AU (plus 140, minus 80), a perihelion of 300 AU (plus 85, minus 60), and an inclination of 16 degrees, give or take 5. They reported the clustering as significant at the 99.6 percent level.

That is a smaller, closer and brighter planet than the 2016 version, and it makes a difference. A body of that size at 300 to 400 AU would reflect enough sunlight to be within reach of the largest ground-based survey telescopes, at roughly magnitude 21 to 23. Its motion against the background stars would be slow, a few arcminutes per year, but measurable across seasons.

The hypothesis is genuinely falsifiable in two ways. Either a telescope finds the planet, or a survey with a well-understood selection function collects enough new distant objects to show whether the alignment is real or an artifact of where and when astronomers happened to look. Both tests are now underway.

Timeline

• 14 November 2003Sedna is discovered by Brown, Trujillo and Rabinowitz; its 76 AU perihelion cannot be explained by Neptune.
• 27 March 2014Trujillo and Sheppard announce 2012 VP113 in Nature and note the clustering of arguments of perihelion among distant bodies.
• 20 January 2016Batygin and Brown publish "Evidence for a Distant Giant Planet in the Solar System" in the Astronomical Journal, coining the name Planet Nine.
• 2016 to 2019: Sheppard, Trujillo and others add new ETNOs; Brown and Batygin publish a review of the hypothesis in Physics Reports.
• 2019 to 2020: Scholtz and Unwin propose that the perturber could be a primordial black hole rather than a planet.
• 2020Sefilian and Touma argue that a massive disc of icy debris could mimic the planet's effect.
• February 2021Napier and colleagues report in the Planetary Science Journal that a bias-corrected sample of 14 ETNOs shows no significant clustering.
• August 2021Brown and Batygin publish "The Orbit of Planet Nine," lowering the estimated mass to about 6 Earth masses and the semi-major axis to about 380 AU.
• 2023Brown and Mathur test Modified Newtonian Dynamics as an alternative and find it does not reproduce the observed pattern.
• March 2024A Pan-STARRS1 search by Brown, Holman and Batygin, combined with earlier Zwicky Transient Facility and Dark Energy Survey limits, rules out 78 percent of the predicted parameter space.
• April 2024Batygin and colleagues argue that a population of low-inclination, Neptune-crossing objects is a further signature of Planet Nine.
• April 2025Terry Long Phan and colleagues report one strong candidate from a comparison of IRAS (1983) and AKARI (2006 to 2011) far-infrared surveys.
• 21 May 2025The Minor Planet Center announces 2017 OF201, a dwarf-planet candidate with a semi-major axis of about 830 AU whose orbit lies outside the cluster.
• 14 July 2025The Subaru Telescope team announces sednoid 2023 KQ14, nicknamed Ammonite, in Nature Astronomy; its orbit points the opposite way from the other sednoids.
• Late February 2026: Rubin Observatory begins issuing public alerts.
• 2 April 2026Rubin reports more than 11,000 new asteroids from a six-week test run, including 380 trans-Neptunian objects.
• 5 June 2026Batygin tells The Debrief that Rubin's discoveries will provide a direct test of every line of evidence for Planet Nine.
• 29 to 30 June 2026: Rubin formally begins the ten-year Legacy Survey of Space and Time.

The Evidence

The primary evidence is dynamical, not photographic. Three separate signatures are claimed.

First, the clustering of ETNO orbits. In the 2016 paper the six objects with semi-major axes above 250 AU had arguments of perihelion grouped within a narrow range and longitudes of perihelion pointing in roughly the same direction. Their orbital poles were also aligned, meaning the orbits tilt the same way. Batygin and Brown calculated the odds of this arrangement arising by chance at 0.007 percent. In the 2021 update, with more objects and an explicit model of observational bias, they put the significance at 99.6 percent.

Second, the perpendicular objects. A handful of bodies orbit the Sun on paths tilted by 90 degrees or more relative to the planets. Ordinary planetary perturbations do not produce these; in the Planet Nine simulations they emerge naturally.

Third, the Neptune-crossers. A 2024 paper led by Batygin argued that a set of objects with low inclinations but orbits that cross Neptune's is difficult to explain without a distant perturber, and offered predictions that new surveys could check.

Against this stand the non-detections. The Pan-STARRS1 search, combined with the Zwicky Transient Facility and the Dark Energy Survey, excluded 78 percent of the 2021 parameter space to a 50 percent completeness depth of visual magnitude 21.5. What remains is mostly fainter than magnitude 21 and concentrated near where the northern Galactic plane crosses the ecliptic, a crowded region where a slow-moving point is easy to lose.

There is one candidate. Phan and colleagues compared far-infrared sources in IRAS and AKARI, whose 23-year separation would let an object at several hundred AU shift by a detectable amount. They found 13 candidate pairs consistent with heliocentric distances of 500 to 700 AU and masses of 7 to 17 Earth masses, and singled out one as strong. Nothing about it has been confirmed; the authors themselves call for follow-up with instruments such as DECam.

Then there are the two objects that do not fit. 2017 OF201, found by Sihao Cheng's team at the Institute for Advanced Study, is roughly 700 kilometres across, currently 90 AU away, with a semi-major axis of about 830 AU and a perihelion of 45 AU. Its longitude of perihelion lies outside the cluster, and the discoverers report that at least one proposed Planet Nine configuration destabilises it quickly. Ammonite, 2023 KQ14, is only the fourth sednoid ever found, with a perihelion of 66 AU, and its orbit is oriented opposite to the other three. The Subaru team's simulations suggest that if a large planet exists it must be farther out than originally proposed.

The numbers, then, are stark: fewer than two dozen relevant orbits, one unconfirmed heat source, and three quarters of the predicted sky already searched.

The Theories

What Doesn't Add Up

• Why has nothing been found in the 78 percent of the predicted sky already searched to magnitude 21.5? Either the planet is fainter, farther, or in the crowded strip near the Galactic plane, or it is not there.
• Why do the two most significant recent discoveries, 2017 OF201 and Ammonite, both fall outside the cluster the planet is supposed to maintain? Two exceptions in a sample of about twenty is a large fraction.
• Which orbital elements are actually clustered? Trujillo and Sheppard flagged the argument of perihelion; Batygin and Brown emphasise the longitude of perihelion and orbital pole; Napier and colleagues find neither survives bias correction in their sample.
• Can the infrared candidate of Phan and colleagues be recovered at all? Its implied distance of 500 to 700 AU is beyond the 2021 best estimate, and no optical counterpart has been reported.
• How did such a planet get there? Proposed origins include ejection from the inner Solar System by Jupiter and capture from a passing star in the Sun's birth cluster. Neither has a decisive test.
• If Rubin finds no planet but confirms the clustering in a bias-controlled sample, what replaces the planet? None of the alternatives currently fits the full set of signatures.

Latest Developments

The decisive instrument has switched on. Rubin Observatory, on Cerro Pachón in Chile, with its 8.4-metre mirror and 3.2-gigapixel camera, began issuing public alerts in late February 2026. On 2 April 2026 the Minor Planet Center confirmed more than 11,000 new asteroids from a six-week engineering run, among them 380 trans-Neptunian objects, close to 8 percent of the previously known population, and two bodies, 2025 LS2 and 2025 MX348, whose orbits reach roughly 1,000 AU from the Sun. The ten-year Legacy Survey of Space and Time formally began on 29 to 30 June 2026.

On 5 June 2026 Batygin told The Debrief that the coming discoveries would be a direct test of all the lines of evidence for Planet Nine. Rubin's uniform cadence lets astronomers inject synthetic objects into the observing history and measure exactly what would have been found, which is the tool the selection-bias debate has lacked. Planetary scientists quoted by National Geographic expect that if the planet lies within the primary search region, it should be identified within the first year or two of operations. As of September 2026 no detection has been announced, and the first population-level statistics are expected with early data releases in 2027.

Sources

1. Caltech Researchers Find Evidence of a Real Ninth Planet — Caltech, 2016
2. Evidence for a Distant Giant Planet in the Solar System — The Astronomical Journal, 2016
3. A Sedna-like body with a perihelion of 80 astronomical units — Nature, 2014
4. The Orbit of Planet Nine — The Astronomical Journal, 2021
5. New Study Knocks Theory of Planet Nine's Existence Out of Orbit — Smithsonian Magazine, 2021
6. A Pan-STARRS1 Search for Planet Nine — The Astronomical Journal, 2024
7. Generation of Low-inclination, Neptune-crossing TNOs by Planet Nine — arXiv / ApJ Letters, 2024
8. A Search for Planet Nine with IRAS and AKARI Data — arXiv / Publications of the Astronomical Society of Australia, 2025
9. Is this the first hint of Planet Nine? — Phys.org, 2025
10. Discovery of a dwarf planet candidate in an extremely wide orbit: 2017 OF201 — arXiv, 2025
11. Subaru Telescope Discovers "Fossil" of the Early Solar System — Subaru Telescope / NAOJ, 2025
12. Early Data from NSF-DOE Vera C. Rubin Observatory Reveals Over 11,000 New Asteroids — Rubin Observatory, 2026
13. Rubin Observatory begins landmark 10-year timelapse of night sky — University of Washington News, 2026
14. A Massive Hidden Planet May Be Lurking Beyond Neptune — The Debrief, 2026
15. Rubin can test Planet Nine by finding the world directly or determining whether the orbital patterns survive — Space Daily, 2026

The Boötes Void

A sphere 330 million light-years wide should hold about 2,000 galaxies. Astronomers have found roughly 60.

The Hook

In 1981 Robert Kirshner and three colleagues were measuring the redshifts of galaxies in three small patches of sky, each about 35 degrees from the others, to build a map of how matter is distributed at large scales. The maths of their survey predicted that the galaxies should pile up at a recession velocity of about 15,000 kilometres per second. Instead, across 133 galaxies, exactly one fell inside the 6,000 kilometres-per-second window centred on that speed.

The same gap appeared in all three fields. The team did what cautious scientists do: they published a paper with a question mark in the title. "A million cubic megaparsec void in Boötes?" appeared in the Astrophysical Journal Letters that year.

The question mark has since been removed. The region is real, roughly spherical, about 330 million light-years across, and centred some 700 million light-years from Earth. A volume that should contain on the order of 2,000 galaxies contains, on the best current count, about 60. Astronomer Greg Aldering, who helped catalogue those galaxies, has remarked that a Milky Way sitting at its centre would not have known other galaxies existed until the 1960s. The void is not a mystery of what is there. It is a mystery of what is not.

The Facts

The discovery paper was written by Kirshner, then at the University of Michigan, with Augustus Oemler Jr., Paul Schechter and Stephen Shectman. Their survey targeted galaxies brighter than roughly R magnitude 16.3 in three fields. The deficiency was consistent in each field between recession velocities of about 12,000 and 18,000 kilometres per second, and on that basis they proposed that the whole region between the fields was empty. The volume of a sphere of radius 62 megaparsecs is very close to one million cubic megaparsecs, which is where the title came from.

A dedicated follow-up survey by the same four authors, published in the Astrophysical Journal in 1987, confirmed the void's existence and dimensions. Modern figures put the radius at about 62 megaparsecs, or nearly 200 million light-years, giving a diameter of roughly 330 million light-years, and the centre at right ascension 14 hours 50 minutes, declination plus 46 degrees, in the constellation of Boötes. The Hercules Superclusters lie on its near edge. It is routinely described as a supervoid, one of the largest known in the observable Universe.

It is not perfectly empty. Beginning in 1987, searches for emission-line galaxies found a handful of objects inside. Infrared-selected IRAS galaxies followed in 1990. A neutral-hydrogen survey with the Very Large Array in the mid-1990s, led by Arpad Szomoru, pointed at sixteen fields around known void galaxies, covering about one percent of the void's volume, and found eighteen previously uncatalogued companion galaxies purely by their radio emission. By 1997, roughly 60 galaxies were known, a figure that has not changed much since. Spectroscopic work in 1997 and 2002 by Shawn Cruzen and colleagues showed that these galaxies are mostly late-type, gas-rich, blue and actively forming stars, and that a few host active galactic nuclei.

The void is not considered a contradiction of the standard cosmological model. Voids in general make up around 80 percent of the volume of the observable Universe and typically span 30 to 300 million light-years. Boötes sits at the top of that range. The favoured account is that it formed by the merging of several smaller voids, which would leave the compressed walls between them as faint internal sheets or filaments. A roughly cylindrical arrangement of galaxies running through the centre of the void has been reported and is often cited as a relic of exactly that process.

Two things make Boötes unusual among voids. First, its sheer size. Second, the fact that its galaxies, though few, include some that are bright and massive, which is not what the emptiest environments are expected to produce.

Timeline

• 1978Gregory and Thompson, and separately Jõeveer and Einasto, report the first large empty regions in galaxy redshift surveys.
• 1981Kirshner, Oemler, Schechter and Shectman publish "A million cubic megaparsec void in Boötes?" in the Astrophysical Journal Letters.
• 1983The first Center for Astrophysics redshift survey is published, showing that voids are a general feature of the galaxy distribution.
• 1987Kirshner and colleagues publish "A survey of the Bootes void," confirming the region with a dedicated redshift survey.
• 1987A search for emission-line galaxies, including work by Moody and Kirshner, reports the first handful of galaxies inside the void.
• 1990Dey, Strauss and Huchra identify IRAS infrared galaxies within the void.
• 1992Weistrop and colleagues report the interacting galaxy pair CG 692 and CG 693 inside the void.
• 1995 to 1996: Szomoru, van Gorkom and Gregg survey neutral hydrogen with the VLA, detecting 16 known galaxies and 18 new companions.
• 1997Cruzen, Weistrop and Hoopes publish spectroscopy of void galaxies; the known population reaches about 60.
• 1997Sage and colleagues detect molecular gas by carbon monoxide emission in several Boötes void galaxies with the IRAM 30-metre telescope.
• 2002Cruzen and colleagues report that some void galaxies host active galactic nuclei.
• 2011Pan and colleagues publish a catalogue of more than a thousand voids from the Sloan Digital Sky Survey, placing Boötes in context.
• 2021Pandey, Saha and Pradhan publish a deep ultraviolet survey of the void with AstroSat's UVIT, finding three previously undetected star-forming galaxies.
• 2025Sharma and colleagues publish the first resolved map of molecular gas in a void galaxy, CG 910, in the Boötes void.
• 2026Tavasoli's study of void evolution across cosmic time reports a population of galaxies near void centres in SDSS data that simulations do not predict.

The Evidence

The evidence for the void is statistical, and the statistics are simple. Kirshner's original survey measured 133 redshifts. The luminosity function of galaxies and the magnitude limit of the survey predicted a peak in the redshift distribution near 15,000 kilometres per second. One galaxy out of 133 appeared in the 6,000 kilometres-per-second band around that peak. The chance of that under a uniform distribution was negligible, and the same deficit appeared in all three widely separated fields.

The scale is set by the redshift range. A recession velocity of 12,000 to 18,000 kilometres per second corresponds to a redshift of roughly 0.04 to 0.06, or a distance of a few hundred megaparsecs. The galaxy CG 910, which sits in the void, has a redshift of 0.045 and a distance of 188 megaparsecs, about 610 million light-years, consistent with the void's quoted centre at 700 million light-years.

The expected galaxy count is a rough number. Using an average density of about one galaxy per 10 million light-years in a region of this volume yields on the order of 2,000 galaxies. Published counts of what has actually been found range from 58 to 60, and the total has been essentially flat for nearly thirty years despite deeper searches. The 2021 UVIT survey covered 615 square arcminutes of the void to limiting magnitudes of 23.0 in the far ultraviolet and 24.0 in the near ultraviolet, and found six star-forming galaxies, three of them new. That is a low yield for a deep search.

The galaxies that are there are informative. Szomoru's radio survey found them gas-rich and similar in properties to ordinary field galaxies, suggesting they formed in small local overdensities within the void rather than by any exotic route. The 2025 study of CG 910 measured a molecular gas mass of about 12 billion solar masses in a regularly rotating disc 7 kilo-parsecs across, with a flat rotation velocity of 256 kilometres per second, plus about 3 billion solar masses of atomic hydrogen. Its star formation rate of 0.33 solar masses per year implies a gas depletion time of tens of billions of years. The authors saw no clear sign of gas accretion or interaction. CG 910 is, in other words, a large, slowly evolving spiral sitting in near-total isolation, with its nearest neighbour 229 kilo-parsecs away.

The spatial arrangement is the last clue. The void's galaxies are not scattered evenly. They trace a roughly tubular structure through the interior, which is the pattern expected if smaller voids merged and left their shared walls behind.

Nothing in this evidence requires new physics. The Lambda-CDM model of cosmology, in which dark energy and cold dark matter drive structure formation from quantum fluctuations in the early Universe, produces voids of this size in simulations. The open questions are about the details.

The Theories

What Doesn't Add Up

• Why has the galaxy count stalled at about 60 since 1997 despite deeper surveys? Either the void really is that empty or the interior has never been systematically imaged to faint limits over its full extent.
• How did a galaxy as massive and gas-rich as CG 910, with 21 billion solar masses of stars and a 12-billion-solar-mass molecular disc, assemble in near-perfect isolation? Standard galaxy growth relies on mergers and accretion that a void largely denies.
• Why are the void's galaxies arranged along a tube through the middle? If it is a relic wall from merged sub-voids, its age and kinematics should be measurable. They have not been.
• Is there a hidden dark-matter or gas component? Voids are not expected to be truly empty of matter, and the baryons that never formed galaxies should be somewhere, likely as diffuse warm gas. No direct detection in Boötes exists.
• Does the 2026 finding of central galaxies in SDSS voids apply to Boötes? If galaxies sit near the centre of the largest voids in numbers simulations do not produce, something in the models of galaxy formation at extreme low density is incomplete.
• What is the void's true shape and boundary? The sphere is an approximation from sparse redshift data. Modern surveys such as DESI could map its walls precisely, but no dedicated modern mapping has been published.

Latest Developments

The most recent direct observations of the void's interior come from two sources. In 2021 a team led by D. Pandey used the Ultraviolet Imaging Telescope on AstroSat to survey 615 square arcminutes of the void to unprecedented ultraviolet depth, cataloguing six star-forming galaxies, three of them new, and finding a bimodal colour distribution among them. In March 2025 Ekta Sharma and colleagues published the first resolved carbon monoxide map of any void galaxy, CG 910, using CARMA, together with Green Bank Telescope hydrogen data. The galaxy turned out to be gas-rich but sluggish, with no sign of the gas accretion that might explain how it grew.

At the population level, a February 2026 study by Saeed Tavasoli compared Millennium Simulation voids with SDSS voids at redshift below 0.04 and found that observed voids become emptier over cosmic time as expected, but also contain galaxies near their centres that the simulations lack. Boötes is among the nearby voids that data from DESI, Euclid and eventually the Rubin Observatory's LSST should be able to map to far fainter limits than Kirshner's 1981 survey. No new galaxy census of the void had been published as of September 2026.

Sources

1. A million cubic megaparsec void in Bootes? — Astrophysical Journal Letters via NASA ADS, 1981
2. A survey of the Bootes void — Astrophysical Journal via NASA ADS, 1987
3. An Historical View: The Discovery of Voids in the Galaxy Distribution — arXiv, 2011
4. An HI survey of the Bootes Void. I. The Data — arXiv / Astronomical Journal, 1996
5. Spectroscopy of Galaxies in the Bootes Void — Astronomical Journal, 2002
6. The Ultraviolet Deep Imaging Survey of Galaxies in the Bootes Void I — arXiv / Astrophysical Journal, 2021
7. The Molecular and Atomic Hydrogen Gas Content of the Boötes Void galaxy CG 910 — arXiv / Astronomy & Astrophysics, 2025
8. Evolution of Cosmic Voids: Structure, Galaxies, and Dynamics — arXiv, 2026
9. Cosmic Voids in Sloan Digital Sky Survey Data Release 7 — arXiv / MNRAS, 2011
10. The Boötes void: Why the Universe has a mysterious hole 330 million light-years across — BBC Science Focus, 2022
11. There's an enormous void of nothingness in our Universe. And scientists found it by accident — BBC Sky at Night Magazine, 2023
12. The Boötes Void is a 'holy grail' of astronomy — Astronomy Magazine
13. Next Stop: Voids — NASA Blueshift, 2013
14. Boötes Void — Wikipedia

The Somerton Man (Tamam Shud)

DNA gave him a name after 74 years. It has not explained the torn page, the code, or how he died.

The Hook

At about 6.30 on the morning of 1 December 1948, a man was found lying against the seawall at Somerton Park beach, south-west of Adelaide, head resting on the concrete, legs extended, feet crossed, a half-smoked cigarette on his collar. He looked like a man who had fallen asleep. He was dead.

He carried no wallet, no hat, no identification. Every label had been cut from his clothes. Months later, during a re-examination of his clothing, a tiny fob pocket was found sewn inside his trousers, and inside it a rolled scrap of paper bearing two printed words: "Tamám Shud." It is Persian, and it means "ended" or "finished." The scrap had been torn from the final page of a copy of the Rubaiyat of Omar Khayyam.

Police found the book. In its back cover, under ultraviolet light, were an unlisted telephone number and five lines of pencilled capital letters that no cryptographer has ever read. The telephone number belonged to a young nurse who lived 400 metres from where he died. She told police she did not know him. When shown a plaster cast of his face, she looked as if she was about to faint.

Seventy-four years later, DNA gave the dead man a name. Nothing else about that night has been explained.

The Facts

The body was found by a couple on the sand opposite the Crippled Children's Home on the corner of The Esplanade and Bickford Terrace, Somerton Park, about 11 kilometres from central Adelaide. Witnesses reported seeing a man in the same position the previous evening, 30 November, between about 7 and 8 pm; one couple saw him raise his right arm and let it drop. Nobody saw his face.

He was about 180 centimetres tall, with grey eyes and fair to ginger hair greying at the temples, and estimated at 40 to 45 years old. He was in excellent physical condition, with strong calf muscles and hands that showed no sign of manual work. His pockets held an unused rail ticket from Adelaide to Henley Beach, a bus ticket, an American-made aluminium comb, chewing gum, cigarettes of one brand in the packet of another, and matches. His dental records matched no one.

The autopsy, performed by pathologist John Dwyer, found the man's last meal was a pasty eaten three to four hours before death, which he placed at about 2 am on 1 December. The organs told a story of poisoning: acute gastric haemorrhage, blood in the stomach, congestion of the liver, kidneys and brain, and a spleen three times normal size. Yet no poison was detected. Dwyer said he was convinced the death was not natural. The body was embalmed on 10 December 1948 after no identification could be made.

On 14 January 1949 staff at Adelaide railway station found an unclaimed brown suitcase, checked into the cloakroom after 11 am on 30 November. Its label had been removed. Inside were clothes, shaving gear, a stencilling brush of the kind used by ships' officers, an electrician's screwdriver, a table knife ground into a blade, a pair of sharpened scissors, and a card of orange Barbour waxed thread of a type not sold in Australia, matching thread used to repair a pocket in the trousers the dead man wore. Three items carried a name: "T. Keane" on a tie, "Keane" on a laundry bag, "Kean" on a singlet. No missing T. Keane was found anywhere in the English-speaking world.

The coroner's inquest, under Thomas Erskine Cleland, opened in December 1948 and was adjourned to 17 June 1949. Pathologist Sir John Burton Cleland, who re-examined the body, noted the man's shoes were freshly polished and that there was no evidence of the vomiting or convulsions poisoning would usually produce, raising the possibility the body had been moved. Professor Cedric Stanton Hicks testified that two drugs, later publicly identified as the cardiac glycosides digitalis and ouabain, could kill in a small oral dose and be undetectable; their names were entered as a sealed exhibit. The coroner said he was prepared to find the man died of a glucoside poison not accidentally administered, but could not say by whose hand. He adjourned without a finding. The inquest was briefly reopened on 14 March 1958 and again adjourned sine die.

Timeline

• 16 November 1905Carl "Charles" Webb is born in Footscray, Melbourne, the youngest of six children of a German-born baker.
• 4 October 1941Webb marries Dorothy "Doff" Robertson, a pharmacist and chiropodist; they live in South Yarra.
• 1943Webb's brother Roy dies as a prisoner of war in Burma; his nephew John Keane dies in the UK serving with the RAAF.
• March 1946Dorothy Webb later states her husband attempted suicide with an overdose of ether.
• April 1947Webb leaves his wife. It is the last documented trace of him.
• 30 November 1948A man checks a brown suitcase into the Adelaide railway station cloakroom after 11 am and buys a ticket to Henley Beach he never uses; witnesses see a man lying at Somerton Park beach that evening.
• 1 December 1948, about 6.30 am: The body is found against the seawall.
• 10 December 1948The body is embalmed; police have never before needed to do this.
• 14 January 1949The unclaimed suitcase is found at the railway station.
• 6 to 14 June 1949: The "Tamám Shud" scrap is discovered in a concealed fob pocket.
• 17 and 21 June 1949: The coroner's inquest hears evidence of poisoning; no finding is recorded.
• 22 July 1949A man hands in a 1941 Whitcombe and Tombs edition of the Rubaiyat with the final words torn out; police find a phone number and pencilled letters in its back cover.
• 26 July 1949The number is traced to Jessica Thomson of Moseley Street, Glenelg, who denies knowing the man but reacts strongly to his plaster cast.
• 27 July 1949Sydney police locate Alf Boxall, to whom Thomson had given a Rubaiyat in 1945; his copy is intact.
• 14 March 1958The inquest is reopened and adjourned indefinitely.
• 1978ABC television's Inside Story has Defence cryptographers examine the code; they cannot decipher it.
• 1986The suitcase and contents are destroyed by police as no longer required.
• 2009Professor Derek Abbott of the University of Adelaide begins his investigation.
• 14 October 2019Attorney-General Vickie Chapman approves exhumation.
• 19 May 2021The body is exhumed from West Terrace Cemetery under Operation Persevere.
• 26 July 2022Abbott and genealogist Colleen Fitzpatrick announce the man is Carl "Charles" Webb.
• 6 July 2026ABC reports that SA Police's report to the coroner is expected imminently, with findings anticipated by the end of 2026.

The Evidence

The physical evidence divides into three sets: the body, the book and the DNA.

The body yielded the poisoning findings above, plus two anatomical rarities later flagged by Abbott's team: a cymba of the ear larger than the cavum, seen in 1 to 2 percent of Europeans, and hypodontia of both lateral incisors, present in about 2 percent of the population. The original 1948 and 1949 autopsy reports are now missing. A plaster cast of the head and shoulders, made after the inquest by technician Paul Lawson, survived, and hairs embedded in it became the case's most important physical exhibit.

The book was a 1941 edition of Edward FitzGerald's translation published by Whitcombe and Tombs in Christchurch, New Zealand. Microscopic examination matched the torn scrap to its last page. The five lines in the back read, as generally transcribed: WRGOABABD, MLIAOI (struck through), WTBIMPANETP, MLIABOAIAQC, ITTMTSAMSTGAB. In 1978 Department of Defence cryptographers reported the text too short to yield a meaning, if it was a cipher at all. A 2014 analysis by computational linguist John Rehling concluded the letters were most likely initials of English words, possibly a private shorthand, and could never be reconstructed. The original book was lost in the 1950s.

The DNA is the newest and strongest evidence. In December 2017 Abbott announced that three hairs suitable for DNA extraction had been recovered from the plaster cast. In 2018 mitochondrial sequencing placed the man in haplogroup H4a1a1a, carried by about 1 percent of Europeans. In 2022 further extraction produced an autosomal profile suitable for forensic genetic genealogy. Abbott and Fitzpatrick built a large family tree from matches on both the paternal and maternal sides, converging on descendants of two distant cousins of one man: Carl Webb.

The circumstantial fit is striking. Webb's sister Freda married a Thomas Keane, and Freda's son John Keane, an RAAF airman who died in 1943, had lived a 20-minute drive from Webb. Family members believe Webb inherited his dead nephew's clothing, which may explain the American tailoring and the Keane name tags. Webb was an electrical instrument maker, which fits the electrician's screwdriver. Dorothy Webb's 1951 divorce affidavit, citing desertion, described a solitary, moody and sometimes violent man who wrote poems about death and had tried to kill himself. By 1951 Dorothy was living in Bute, South Australia, 144 kilometres from Adelaide.

Abbott has said he is 99.99 percent confident. The formal chain, however, is unfinished. Forensic Science SA holds the exhumed remains and has not published its own DNA result. South Australia Police described itself in July 2022 as cautiously optimistic and in July 2026 as still conducting a complex and ongoing coronial investigation.

The Theories

What Doesn't Add Up

• Why Adelaide, and why Glenelg? Webb's estranged wife was later recorded in Bute, but that was 1951; no evidence places her there in 1948. His bus ticket took him to within a kilometre of Jessica Thomson's house.
• Why was her unlisted telephone number in his book? Thomson lied to police about her marital status, asked that her name be withheld, told her daughter she had lied about knowing him, and was described by Detective Gerry Feltus, who interviewed her in 2002, as evasive. She died in 2007 without explaining.
• What are the five lines of letters? The frequency analysis says they are not random and not a substitution cipher. The best modern guess is initials, possibly of horse names, since Webb bet on races. That is a guess.
• Was the body moved? Freshly polished shoes after a day walking around Glenelg, no vomit, no convulsion marks, a witness who saw a man carrying another, and a coroner who said the difficulties would vanish if the body had been carried.
• What was the poison? Never detected. The 1948 pathology suggested a barbiturate; the 1949 inquest suggested digitalis or ouabain; the Chief Justice of Victoria, John Harber Phillips, concluded digitalis in 1994. None has been confirmed, and Abbott says the skeletal remains cannot now reveal it.
• Why was the label on the suitcase removed, and why no socks? A man who packs pyjamas and slippers but no spare socks or correspondence, and strips the tag from his own luggage, was either hiding something or being hidden.
• Where is the book? The original Rubaiyat, the only primary source for the code, was lost by police in the 1950s. The suitcase was destroyed in 1986.

Latest Developments

The case has moved from identification to formalisation, slowly. In November 2022 the ABC's Australian Story broadcast photographs from a Webb family album showing a man named Charlie in the 1920s whose features match the 1948 autopsy photographs, and reported Dorothy Webb's divorce affidavit describing his suicide attempt and violent moods. Webb's relatives, none of whom knew him, accepted the identification.

On 6 July 2026 the ABC reported that, five years after the exhumation, SA Police had still not delivered its report to the coroner. Abbott said police had told him it would be complete within two weeks and that the coroner was expected to hand down findings by the end of 2026. A police spokesperson would say only that the investigation was complex and ongoing. The remains are believed to be held at the Forensic Science Centre in central Adelaide, and the grave at West Terrace Cemetery, once marked "the unknown man," is empty. Stuart Webb, believed to be the great-grandson of Carl Webb's brother, said the family would probably leave him in Adelaide once released. Abbott's own assessment of the remaining questions is that the cause of death and the reason for the journey will probably never be known.

Sources

1. SA Police report to coroner into Somerton Man case complete 'very soon' — ABC News, 2026
2. Somerton Man identified as Melbourne electrical engineer, researcher says — ABC News, 2022
3. The Somerton Man has been named. What do we know about Carl 'Charles' Webb? — ABC News, 2022
4. What do we know about Carl Webb? Researchers find ads, photographs to paint picture of well-to-do man — ABC News, 2022
5. Somerton Man Charles Webb's true identity revealed in family photographs and divorce papers — ABC News, 2022
6. Complete remains of Somerton Man in 'reasonable condition' after exhumation — ABC News, 2021
7. 'Tamam Shud' and the five other clues that have failed to solve the Somerton Man mystery — ABC News, 2021
8. How genetic genealogy is solving Australia's coldest cases — ABC News, 2023
9. Somerton man mystery 'solved' as DNA points to man's identity, professor claims — CNN, 2022
10. Australia's Mysterious Somerton Man Identified as Carl Webb — Smithsonian Magazine, 2022
11. Statement regarding Somerton Man — South Australia Police (archived), 2022
12. 'But what poison?' Mystery of the Somerton Man — State Records of South Australia
13. For 70 years the identity of the Somerton man has been a mystery. His body is now being exhumed — SBS News, 2021
14. "Tamam Shud", The Advertiser, 10 June 1949 — Trove / National Library of Australia, 1949
15. Somerton Man — Wikipedia

The CMB Cold Spot

A patch of the universe's oldest light is colder than it should be. Twenty years of surveys cannot say why.

The Hook

The cosmic microwave background is the oldest light there is: radiation released roughly 380,000 years after the Big Bang, now cooled to 2.725 kelvin and filling every direction of the sky. Its temperature is almost perfectly even. The typical fluctuation from one point to another is about 18 millionths of a degree.

In the southern sky, toward the faint river-constellation Eridanus, there is a region roughly ten degrees across, twenty full Moons side by side, where the temperature drops by around 70 millionths of a degree on average, and deeper still at its core. It is ringed by a band of slightly warmer sky, which makes it harder, not easier, to write off as chance. A standard, purely random universe produces a feature like this about one time in a hundred, or less, depending on how you ask the question.

Astronomers have looked for something in front of it. They found a hole in the cosmic web nearly two billion light-years wide. It is not enough. The math says the void should chill the light by a fraction of what is observed. What makes up the rest is still unknown.

The Facts

The Cold Spot was identified in 2004 by Patricio Vielva and colleagues at the Instituto de Física de Cantabria, working on the first-year all-sky temperature map from NASA's Wilkinson Microwave Anisotropy Probe (WMAP). They did not go looking for a spot. They were running a statistical test using a filter called the Spherical Mexican Hat Wavelet, and noticed an excess of kurtosis, a fat-tailed distribution of values, at angular scales of around 300 arcminutes in the southern hemisphere. Follow-up work by Marcos Cruz and collaborators in 2005 localised the source to galactic coordinates l = 209°, b = -57°, in Eridanus.

The feature is a large-scale temperature decrement, roughly 5° in radius, whose average deficit is about 70 microkelvin relative to the CMB mean. Fitted profiles used in later void studies quote a central depth of order 150 microkelvin. Surrounding the cold core is a ring of slightly hotter-than-average sky. Both the depth and the ring are important: taken together they give the Cold Spot a shape that ordinary Gaussian fluctuations rarely produce.

WMAP's successor, the European Space Agency's Planck satellite, re-observed the whole sky at higher resolution and with independent detectors. In its 2013 results on isotropy and statistics, the Planck collaboration confirmed that the Cold Spot is anomalous in terms of the area of the wavelet coefficients on scales of about 10°, and stated that many anomalies previously seen in WMAP data were reproduced at similar significance, around 3 sigma. The 2015 Planck release repeated the analysis with the same conclusion. Whatever the Cold Spot is, it is not an instrumental artefact of one spacecraft.

Its statistical significance is genuinely disputed. Zhang and Huterer, in a 2010 paper titled "Disks in the sky," repeated the search using simple circular disk filters and Gaussian filters of varying width instead of the wavelet and found no significant signal at any scale. They argued the original significance came at least partly from a fortuitous match between the Cold Spot's profile and the chosen wavelet shape. Planck's later assessment still puts the probability of such a feature under standard cosmology at roughly 1 percent, but the "look-elsewhere" problem, the fact that in a whole sky you will always find something odd if you search hard enough, is a live objection.

In 2015 a team led by István Szapudi at the University of Hawaii reported a supervoid, a vast region of below-average galaxy density, aligned with the Cold Spot. That discovery, and the argument over whether it explains anything, has driven the case ever since.

Timeline

• 2003NASA's WMAP releases its first-year all-sky CMB temperature map.
• 2004Vielva et al. detect a non-Gaussian cold feature in the southern sky using Spherical Mexican Hat Wavelets.
• 2005Cruz et al. localise the Cold Spot to galactic coordinates (209°, -57°) and quantify its size and depth.
• 2007Cruz, Turok, Vielva, Martínez-González and Hobson publish in Science that the spot is consistent with a cosmic texture, a relic defect from the early universe.
• 2010Zhang and Huterer argue that the spot's significance depends on the choice of filter and may be a statistical fluke.
• 2013ESA's Planck collaboration confirms the Cold Spot as anomalous in its first isotropy and statistics paper.
• 2014Nadathur et al. calculate that even an extreme supervoid produces only about -20 microkelvin via the integrated Sachs-Wolfe effect, far short of the observed profile.
• 2015Szapudi et al. report a supervoid roughly 1.8 billion light-years across, about 3 billion light-years away, aligned with the Cold Spot, using WISE and Pan-STARRS1 data.
• 2015Planck's second data release again finds the Cold Spot anomalous.
• 2017Mackenzie et al. (Durham) publish the 2dF-VST ATLAS Cold Spot redshift survey: about 7,000 galaxies at z < 0.4; voids found, but too small to explain the spot.
• 2021/2022 — Kovács et al. use Dark Energy Survey Year-3 data to confirm the Eridanus supervoid at z < 0.2, but find its standard-model imprint covers only 10-20 percent of the deficit.
• 2022Owusu et al. analyse Planck lensing data and report that no void anywhere along the line of sight can be responsible.
• 2023Luparello et al. and Hansen et al. report a temperature decrement in Planck and WMAP maps around nearby spiral galaxies, and propose an unknown foreground.
• 2024Lambas, Hansen et al. show in Astronomy & Astrophysics that a foreground modelled on the Eridanus galaxy group complex reproduces the Cold Spot's position and shape.
• 2025Hansen et al. report a p < 0.0001 detection of CMB cooling in galactic halos, keeping the foreground hypothesis alive.

The Evidence

The primary evidence is two independent all-sky maps. WMAP, launched in 2001, mapped the CMB at five frequencies with an angular resolution of a fraction of a degree. Planck, launched in 2009, observed at nine frequencies with better resolution and sensitivity. Both show the same cold feature at the same location with the same shape. The Planck team tested it against different component-separation methods, sky masks and frequency channels and found it robust to all of them.

The numbers matter. The CMB's root-mean-square temperature variation is about 18 microkelvin. The Cold Spot's average deficit over its 5° radius is about 70 microkelvin, and profile fits used in the void literature reach roughly 150 microkelvin at the centre. The hot ring around it is a distinct signature; it is one reason the wavelet analysis flagged the region so strongly.

The second body of evidence is the galaxy surveys. Szapudi's 2015 detection used infrared data from NASA's WISE satellite and optical data from the Pan-STARRS1 telescope to count galaxies along the line of sight. The team found a large underdensity, about 20 percent below the cosmic average, centred at a redshift near 0.2 and extending across roughly 1.8 billion light-years. The 2017 Durham survey sampled about 7,000 galaxies across the inner 5° of the spot with the Anglo-Australian Telescope's 2dF spectrograph and found underdense regions at redshifts 0.14, 0.26 and 0.30, interleaved with small overdensities. The 2022 Dark Energy Survey analysis, using the redMaGiC catalogue of luminous red galaxies, confirmed the Eridanus supervoid as a significant underdensity at z < 0.2 and described it as the dominant underdensity in the local cosmic neighbourhood.

The third line is the physics linking voids to the CMB: the integrated Sachs-Wolfe (ISW) effect. In a universe dominated by dark energy, a photon crossing a large void loses a little energy because the void's gravitational potential decays while the photon is inside it. The effect is real and has been detected statistically across many voids. But for one void to make the Cold Spot, the calculations by Nadathur et al. in 2014 and the DES team in 2022 agree: the standard-model imprint is around 20 microkelvin, or 10-20 percent of what is seen. The DES paper also notes an unresolved oddity: stacked supervoids elsewhere in the sky show ISW imprints about 5.2 ± 1.6 times stronger than the standard model predicts.

The fourth line is gravitational lensing. A large void bends the CMB in a characteristic way. Owusu and collaborators searched the Planck lensing map for that signature and reported none consistent with a void large enough to produce the spot, anywhere between Earth and the surface of last scattering.

The Theories

What Doesn't Add Up

• Why is the largest known local underdensity aligned with the sky's most anomalous cold region, if the void explains only a fifth of the effect? Coincidence is possible, but the alignment is precise and the void is exceptional.
• Why do stacked supervoids elsewhere on the sky show ISW imprints roughly five times stronger than standard cosmology predicts, as the DES team reported in 2022? If that excess is real, the Eridanus void might do more than the models allow, and the models are what need fixing.
• Why does the Cold Spot come with a hot ring? Random fluctuations produce cold patches, but the ring is a specific morphology that makes both chance and simple void models less comfortable.
• How significant is it, really? Wavelet analyses say 1 percent or better; disk-filter analyses say nothing unusual. Two decades on, the field still has not agreed on the correct way to ask the question.
• If the 2023-2025 foreground signal around nearby galaxies is real, what physical process cools microwave photons over several megaparsecs around spiral galaxies? No known mechanism does this at the required strength.
• Why has no confirmed polarisation signature been found that would settle the texture hypothesis one way or the other?

Latest Developments

The most active thread since 2023 has been the foreground hypothesis. Luparello et al. (2023) reported the galaxy-halo temperature decrement; Hansen et al. argued in the same year that it could account for several CMB anomalies at once; and the January 2024 Astronomy & Astrophysics paper by Lambas, Hansen and colleagues explicitly modelled the Eridanus galaxy group complex and recovered a Cold Spot look-alike at the right place. In April 2025 Hansen's group published a further A&A paper claiming a p < 0.0001 detection of CMB cooling in galactic halos and speculating on a link to dark matter. These results are not yet independently confirmed, and the physical mechanism is unknown.

On the void side, the December 2022 lensing analysis by Owusu et al. remains the strongest argument against any supervoid explanation, while the Dark Energy Survey's 2022 confirmation of the Eridanus underdensity, and its note of an unexplained ISW excess in other supervoids, keeps the question of whether standard cosmology is mis-modelling large voids on the table. No published result in 2025 or 2026 has closed the case in either direction.

Sources

1. Planck 2013 results. XXIII. Isotropy and statistics of the CMB — Astronomy & Astrophysics, 2014
2. Planck 2015 results. XVI. Isotropy and statistics of the CMB — Astronomy & Astrophysics, 2016
3. A Cosmic Microwave Background Feature Consistent with a Cosmic Texture (Cruz et al.) — Science, 2007
4. Disks in the sky: A reassessment of the WMAP "cold spot" (Zhang & Huterer) — Astroparticle Physics, 2010
5. Can a supervoid explain the Cold Spot? (Nadathur et al.) — Physical Review D, 2014
6. Cold cosmic mystery solved: Largest known structure in the universe leaves its imprint on CMB radiation — Phys.org / University of Hawaii, 2015
7. Evidence against a supervoid causing the CMB Cold Spot (Mackenzie et al.) — Monthly Notices of the Royal Astronomical Society, 2017
8. Competing claims over cause of cosmic cold spot — Physics World, 2017
9. DES view of the Eridanus supervoid and the CMB cold spot (Kovács et al.) — Monthly Notices of the Royal Astronomical Society, 2022
10. Scientists move a step closer to understanding the "cold spot" in the cosmic microwave background — Fermilab News, 2022
11. The CMB cold spot under the lens: ruling out a supervoid interpretation (Owusu et al.) — arXiv, 2022
12. Perhaps a supervoid doesn't explain the mysterious CMB cold spot — Phys.org, 2022
13. The CMB Cold Spot as predicted by foregrounds around nearby galaxies — Astronomy & Astrophysics, 2024
14. A p < 0.0001 detection of cosmic microwave background cooling in galactic halos and its possible relation to dark matter — Astronomy & Astrophysics, 2025
15. Why a giant 'cold spot' in the cosmic microwave background has long perplexed astronomers — Space.com

The Bermuda Triangle

Six aircraft and 27 men vanished in one afternoon in 1945. The Navy's final verdict was "cause unknown."

The Hook

Shortly after 2:00 p.m. on 5 December 1945, five TBM Avenger torpedo bombers lifted off from Naval Air Station Fort Lauderdale for a routine training exercise called Navigation Problem No. 1: fly east, drop practice bombs on Hens and Chickens Shoals, turn north over Grand Bahama, then home. The flight leader, Lieutenant Charles C. Taylor, had roughly 2,500 hours in the air and combat tours in the Pacific. Around 3:45 p.m. the tower heard him say he could not see land and believed his compasses had failed. Over the next several hours the radio traffic grew more confused, the weather closed in, and darkness fell. The aircraft ran out of fuel somewhere over the Atlantic. A PBM-5 Mariner flying boat with 13 men aboard took off from NAS Banana River to search for them and was never heard from again. Five days and 250,000 square miles of searching produced no wreckage, no raft, not even an oil slick.

Fourteen men on Flight 19, thirteen on the Mariner. The Navy's investigation first blamed Taylor's disorientation, then, after his mother protested, amended the finding to "cause unknown." That phrase built an industry. Whether it deserved to is the question this file tries to answer.

The Facts

The "Bermuda Triangle" is not a place in any official sense. NOAA's National Ocean Service states that no official map delineates its boundaries and the U.S. Board on Geographic Names does not recognise the name. The loosely agreed vertices are Miami, Bermuda and Puerto Rico. The term was coined by writer Vincent Gaddis in a cover story, "The Deadly Bermuda Triangle," in the February 1964 issue of the pulp magazine Argosy, which claimed more than 1,000 lives lost in the region. Charles Berlitz's 1974 book The Bermuda Triangle sold on the order of 20 million copies in 30 languages and made the legend global. A year later, librarian and pilot Larry Kusche published The Bermuda Triangle Mystery: Solved, which went back to the original newspaper and official records for each case and found that many had been misreported, some had happened far outside the region, and at least one ship cited did not exist.

The official position of the agencies responsible for the water is unambiguous. The U.S. Coast Guard states that it does not recognise the existence of the so-called Bermuda Triangle as a geographic area of specific hazard to ships or planes, and that reviews of losses have found nothing to indicate causes other than physical ones. NOAA says there is no evidence that mysterious disappearances occur with any greater frequency there than in any other large, well-travelled area of ocean. Lloyd's of London, according to accounts collected by sceptics such as Australian science broadcaster Karl Kruszelnicki, found no evidence of excess losses and does not charge higher premiums for passage through the area.

Set against that are the cases that will not quite go away. The U.S. Navy collier USS Cyclops, 540 feet long, left Barbados on 4 March 1918 for Baltimore with a cargo of manganese ore and 306 people aboard and was never seen again; no distress call was received and no wreck has been found. It remains the largest non-combat loss of life in U.S. naval history. Two British South American Airways Avro Tudor airliners, Star Tiger and Star Ariel, vanished within a year of each other, in January 1948 and January 1949, with 31 and 20 people aboard respectively; the British inquiry into Star Tiger wrote that no more baffling problem had ever been presented for investigation. The tanker SS Marine Sulphur Queen, carrying molten sulphur and 39 crew, disappeared on 4 February 1963 in or near the Straits of Florida; a Coast Guard Marine Board of Investigation could not fix a cause but concluded the ship should never have sailed.

Each of these has a documented, mundane candidate explanation. None has a proven one.

Timeline

• 4 March 1918USS Cyclops departs Barbados for Baltimore with 306 aboard; due 13 March; never arrives, no distress signal.
• 23-25 November 1941 — USS Proteus, sister ship of Cyclops, lost with all 58 crew in the Caribbean, cause unknown.
• 10 December 1941USS Nereus, second sister ship, lost after leaving St. Thomas, Virgin Islands, on the same route.
• 5 December 1945Flight 19 (five TBM Avengers, 14 men) and a PBM-5 Mariner rescue aircraft (13 men) lost off Florida.
• 1945-1947Navy Board of Investigation blames flight leader's disorientation; finding later amended to "cause unknown."
• 30 January 1948BSAA Star Tiger vanishes between the Azores and Bermuda with 31 aboard.
• 17 January 1949BSAA Star Ariel vanishes between Bermuda and Kingston, Jamaica, with 20 aboard; Avro Tudor withdrawn from passenger service.
• 2 February 1963SS Marine Sulphur Queen sails from Beaumont, Texas, with 39 crew; last routine radio contact 4 February.
• 20 February - 13 March 1963 — Navy and Coast Guard recover life jackets, life rings, name boards and a fog horn near Key West; no hull, no bodies.
• February 1964Vincent Gaddis names "The Deadly Bermuda Triangle" in Argosy.
• 17 March 1964Coast Guard Marine Board publishes findings on the Marine Sulphur Queen: four possible causes, all design or maintenance related.
• 1974Charles Berlitz's The Bermuda Triangle becomes an international bestseller.
• 1975Larry Kusche's The Bermuda Triangle Mystery: Solved dismantles the case files one by one.
• 2010NOAA posts its Ocean Facts statement: no evidence of unusual disappearance rates.
• 2018University of Southampton oceanographer Simon Boxall models rogue waves against a scale USS Cyclops for a Channel 5 documentary.
• January 2020Wreck 35 miles off St. Augustine, Florida, identified as SS Cotopaxi, lost with 32 aboard in 1925 and long listed as a "Triangle" victim; it lies outside the Triangle's usual boundaries.
• December 202580th anniversary of Flight 19; the NAS Fort Lauderdale Museum maintains the memorial.

The Evidence

The evidentiary record of the Bermuda Triangle is a record of absences, which is why it must be handled carefully.

Flight 19. The Naval History and Heritage Command confirms the departure shortly after 2:00 p.m., radio traffic indicating compass failure and disorientation, deteriorating weather, and a five-day search of more than 250,000 square miles that found nothing. NHHC director Samuel Cox's H-gram adds that Taylor had 2,509.3 hours total, 606 in the Avenger, and that the Mariner lost with 13 men, commanded by Lieutenant (j.g.) Walter Jeffrey, made its last radio call at about 7:30 p.m. The tanker SS Gaines Mills reported that at 7:50 p.m. it saw a burst of flames, apparently an explosion, 100 feet high and burning for ten minutes. The PBM was prone to fuel-vapour accumulation and nicknamed the "flying gas tank." The Board of Investigation concluded that Taylor believed he was over the Florida Keys when he was over the Bahamas, and led the flight the wrong way. The dramatic quotations in popular accounts ("everything looks strange, even the ocean") are not reliably sourced to the Board's transcripts; sceptics led by Kusche argue they were embellished afterward.

USS Cyclops. Navy records show a Proteus-class collier launched in 1910, loaded with manganese ore in Brazil, her starboard engine disabled by a cracked cylinder, a defect her commanding officer, Lieutenant Commander George Worley, had reported. Contemporary accounts put her cargo at 10,000-11,000 tons against a designed capacity of roughly 8,000. Both sister ships, Proteus and Nereus, were lost in the same waters in late 1941, also carrying heavy ore, also without trace. The Navy's summary: wreck never found, cause unknown.

Star Tiger and Star Ariel. Both were Avro Tudor IV airliners of British South American Airways. Star Tiger's cabin heater had failed, and the captain flew the Azores-Bermuda leg at 2,000 feet, which increased fuel burn and left almost no margin for error; the last routine position report came in the early hours of 30 January 1948. The Ministry of Civil Aviation could establish no cause. Star Ariel disappeared in clear weather about an hour out of Bermuda on 17 January 1949, having reported nothing wrong. No wreckage from either aircraft was ever recovered; the Tudor was withdrawn from passenger routes.

Marine Sulphur Queen. This is the one case with physical evidence. The vessel was a 523-foot wartime T2 tanker converted to carry molten sulphur in heated tanks. She left Beaumont on 2 February 1963 with 39 crew. On 20 February a Navy vessel recovered a stencilled life preserver and a fog horn cone about twelve nautical miles southwest of Key West; the search eventually recovered 8 life jackets, 5 life rings, 2 name boards, a shirt, part of an oar, two cans, a cone buoy and the fog horn. The Coast Guard Marine Board's findings, published 17 March 1964, stated that the exact cause could not be ascertained but that the ship was not seaworthy, and listed four scenarios: explosion in the cargo tanks, structural failure of the weak-backed T2 hull, capsizing in heavy seas, or a steam explosion from sulphur leaking into void spaces. Federal courts later held the owners liable.

The Theories

What Doesn't Add Up

• Why did USS Cyclops send no distress signal? She carried radio and was on a known route; even a sudden structural failure usually leaves time for something. Her two sister ships also vanished silently. Three identical hulls, three silent losses, is consistent with a design flaw under ore loading, but that has never been proven.
• Why did the five-day search for Flight 19 find nothing at all? Six aircraft went into the water. Avengers sank quickly and the sea was rough, but the total absence of floating debris across a quarter-million square miles is unusual.
• What happened to Star Ariel? Star Tiger at least had a fuel-and-altitude story. Star Ariel was in clear weather, an hour into the flight, with no reported fault, and simply stopped transmitting.
• Why does the Marine Sulphur Queen's debris field consist almost entirely of loose deck items and no bodies, when 39 men were aboard? Sharks, currents and the speed of a hull failure are all plausible, but the Board could not choose among four different mechanisms.
• If Lloyd's and the Coast Guard are right that the loss rate is ordinary, why do the unsolved cases cluster so tightly in time, 1945, 1948, 1949, 1963? Unlucky clustering is compatible with an ordinary base rate, but it is what keeps the legend alive.
• The archival record for the earliest cases is thin. Kusche found many were misreported; it is not always clear how much of what is "known" about Cyclops is contemporaneous and how much is retelling.

Latest Developments

Nothing in the last three years has changed the agencies' position. NOAA's Ocean Facts page, last updated in June 2024, retains its 2010 statement that there is no evidence of unusual disappearance rates. The most recent surge of coverage came in August 2025, when Karl Kruszelnicki's long-standing argument, that losses are proportionate to heavy traffic and that neither the Coast Guard nor Lloyd's of London sees an anomaly, was recirculated across U.S. news outlets; he had made the same case in 2017, 2022 and 2023.

The most substantive recent physical finding remains the 2020 identification of the SS Cotopaxi wreck, 35 miles off St. Augustine, by diver Michael Barnette after roughly 15 years of research, attributing the 1925 loss to a storm and a ship ill-equipped for it. December 2025 marked the 80th anniversary of Flight 19; the Naval Air Station Fort Lauderdale Museum maintains the memorial, and no wreckage attributable to the five Avengers or the Mariner has been confirmed. USS Cyclops's wreck has not been found. No new inquiry into any of the core cases is known to be open.

Sources

1. The Loss of Flight 19 — Naval History and Heritage Command, 2024
2. H-057-4: The Disappearance of Flight 19 — Naval History and Heritage Command, 2021
3. The Mysterious Disappearance of Flight 19 — Naval History Magazine, U.S. Naval Institute, 2021
4. USS Cyclops Mystery — National Museum of the U.S. Navy / NHHC
5. The Mysterious Fate of the Cyclops — Naval History Magazine, U.S. Naval Institute, 2020
6. "...her loss remains unknown...": The Mysterious Disappearance of the USS Cyclops (AC-4) — U.S. National Archives, 2021
7. Ministry of Civil Aviation report on the loss of Tudor IV G-AHNP Star Tiger — Aviation Safety Network archive, 1948
8. Investigations: The Queen with the Weak Back — TIME, 1964
9. In Re Marine Sulphur Transport Corp., 312 F. Supp. 1081 (S.D.N.Y. 1970) — Justia, 1970
10. What is the Bermuda Triangle? — NOAA National Ocean Service, 2010 (updated 2024)
11. The Bermuda Triangle Mystery Delusion: Looking Back after Forty Years — Skeptical Inquirer, 2015
12. Could Gas Explosions Explain Bermuda Triangle Mystery? — Live Science, 2014
13. Bermuda Triangle mystery solved? — Marine Industry News, 2018
14. A ship was believed to have vanished in the Bermuda Triangle. Scientists found it after almost 95 years — CNN, 2020
15. Is the Bermuda Triangle mystery finally solved? One scientist thinks so — FOX 29 Philadelphia, 2025

The 52-Hertz Whale

A Cold War sub-hunting network heard a whale singing at the wrong pitch. Tracked for decades, never seen.

The Hook

In 1989, at Naval Air Station Whidbey Island in Washington State, a Navy acoustician named Joe George was listening to the ocean through SOSUS, the Sound Surveillance System, a network of hydrophones laid across the Pacific floor to track Soviet submarines. He heard something that was not a submarine. It had the rhythm and structure of a baleen whale call: repeated tones, five to seven seconds long, produced in groups. But it was pitched at 52 hertz. Blue whales, the closest match in pattern, sing mostly between 10 and 40 hertz, with their strongest tones near 15 to 25. Fin whales pulse at about 20. Nothing in the catalogue sang at 52.

When the Cold War ended and SOSUS was partly declassified, researchers at Woods Hole Oceanographic Institution began following the signal. They found only one source. Every season, it appeared somewhere in the North Pacific between August and December, moved tens of kilometres a day along routes that looked like a blue whale's, and faded out of range in January or February. In one season it was tracked over 11,062 kilometres. In more than thirty years of listening, no one has ever seen the animal that makes the sound.

The Facts

The scientific record begins with a paper in Deep-Sea Research I, volume 51 (2004), pages 1889-1901, titled "Twelve years of tracking 52-Hz whale calls from a unique source in the North Pacific," by William A. Watkins, Mary Ann Daher, Joseph E. George and David Rodriguez of Woods Hole Oceanographic Institution. Watkins, who with William Schevill had pioneered marine mammal acoustics at WHOI in the 1950s, died in September 2004, weeks before the paper appeared in print.

The paper reports that distinctive tonal calls with emphasis at 52 Hz were first detected in 1989. For the next three years they were received for a few weeks each year from a single source that stayed generally near 46°N, 126°W, off the coast of Oregon and Washington. From 1992 onward, using SOSUS and other Navy hydrophone arrays, the team tracked the source every season through 2004. NOAA's Pacific Marine Environmental Laboratory, which monitors the same arrays, describes the signal as tones with an average centre frequency of 51.75 Hz, lasting 5 to 7 seconds, usually produced in groups of 2 to 6, recorded in the northeast Pacific.

Key numbers from the WHOI work: the call was picked up each year between August and December and followed until the animal moved out of range, always within a few weeks in January or early February. It travelled 31 to 69 kilometres per day. Seasonal tracks ranged from a low of 708 kilometres to a high of 11,062 kilometres in 2002-03. Its range extended from offshore California to the Aleutian Islands. The authors wrote that it was hard to accept that only one animal of this kind could exist across so large an ocean, yet careful year-round monitoring had found only one call with these characteristics anywhere, and only one source each season.

The species is unknown. The call's pattern and seasonality resemble a blue whale's; the timing of some movements has been compared to fin whales. Watkins and colleagues considered a malformed individual or a hybrid of two species. Kate Stafford of NOAA's National Marine Mammal Laboratory, who also monitored the call, has said the seasonality and migratory pattern match blue whales closely enough that she is confident at least part of the animal is a blue whale. The animal's sex is unknown; the widespread use of "he" is convention, not data.

The "loneliest whale" framing is journalistic. Christopher Clark of Cornell University, who recorded the whale in 1993, has said that blue, fin and humpback whales can all hear it and that it sings with many of the features of a typical blue whale song; it is odd, not deaf and not mute. Whether any other whale has ever answered is unknown, because no one has been able to watch.

Timeline

• 1989Navy acoustician Joe George detects the 52 Hz call on SOSUS hydrophones at NAS Whidbey Island; WHOI records it again in 1990 and 1991.
• 1991NOAA PMEL and the National Marine Mammal Laboratory begin joint acoustic monitoring of North Pacific whales using SOSUS.
• 1992With SOSUS partly declassified, WHOI begins systematic seasonal tracking of the single 52 Hz source.
• 1993Cornell's Christopher Clark independently records the call.
• 1995Cascadia Research Collective first documents a blue-fin hybrid whale off California, catalogued as CRC-BM-3330 / CRC-BP-114, later the leading suspect.
• 2002-03Longest single-season track: 11,062 kilometres.
• March-August 2004 — Watkins et al. paper received, accepted and published online in Deep-Sea Research I; Watkins dies in September.
• December 2004The New York Times reports on the study, launching the "loneliest whale" story worldwide.
• 2010Sensors off California record calls matching the 52 Hz pattern on widely separated hydrophones; Scripps oceanographer John Hildebrand suggests more than one animal may be singing.
• 2014-2015Filmmaker Joshua Zeman raises more than $405,000 from over 3,000 Kickstarter backers to fund a search.
• October 2015Seven-day expedition in the Santa Barbara Channel with John Calambokidis (Cascadia), Ana Širović (Texas A&M) and John Hildebrand (Scripps), using sonobuoys, tags and drones; the 52 Hz whale is not found.
• 2015Oregon State University tags the hybrid CRC-BM-3330 for 28.1 days; attempts to match its position with Scripps recordings of the 52 Hz call are inconclusive.
• 2016Calambokidis films the hybrid feeding among 15-25 blue whales and pairing with one of them.
• 2021Jefferson et al. publish the hybrid's 25-year sighting record in Aquatic Mammals; The Loneliest Whale: The Search for 52 is released by Bleecker Street on 9 July.
• 2024Genome study in Conservation Genetics finds an average 3.5 percent fin whale DNA in North Atlantic blue whales, proving hybrids are fertile.
• January 2025Calambokidis and Joe George meet at an Orca Network event and compare notes 35 years after the first detection.
• October 2025The documentary is added to Netflix.
• 2026Pampoulie et al. publish a global review in Ecology and Evolution: 46 possible blue-fin hybrids on record, 17 genetically confirmed.

The Evidence

The primary evidence is acoustic, and it is unusually good for a mystery. SOSUS was built to localise submarines by timing the arrival of sound at multiple hydrophones, which is exactly what a biologist needs to track a calling whale. The WHOI team did not simply hear the call; they fixed its position repeatedly across whole seasons and drew its track on a map. The tracks are long, coherent and consistent with one large animal swimming at whale speeds: 31 to 69 kilometres per day, up to 11,062 kilometres in a season. Ordinary acoustic tracks of individual whales, the paper notes, last hours at best.

The signal itself is well characterised. NOAA PMEL's published spectrogram shows tones averaging 51.75 Hz, 5 to 7 seconds long, in groups of 2 to 6. The structure, repeated tonal units in a patterned sequence, is that of blue whale song, shifted far upward. Reports since the 1990s indicate the call has slowly deepened to around 47 to 49 Hz; this is often read as the animal maturing, though blue whales worldwide have also been documented lowering their pitch over the same decades, for reasons still debated.

The second body of evidence is negative. Across the North Pacific hydrophone network, monitored year-round by WHOI, NOAA PMEL and the Navy, the 2004 paper found no other call with the same characteristics and only one source per season. That is the basis for treating it as a single individual. The 2010 California detections complicate this: calls fitting the Watkins pattern showed up on widely separated Scripps sensors at times suggesting more than one animal. Cascadia Research notes these calls are slightly lower in pitch than the original. Whether they came from the same whale, a second individual, or ordinary blue whales at the high end of their range is unresolved.

The third body of evidence is the hybrid. Cascadia's catalogue whale CRC-BM-3330 / CRC-BP-114 has been sighted 20 times between 1995 and 2020 off California; it has a larger dorsal fin and fainter mottling than a blue whale, consistent with fin whale ancestry, and carried a satellite tag for 28.1 days in 2015. It is a real, photographed blue-fin hybrid living where the 52 Hz call has been heard. What is missing is a recording of it making the call; the 2015 attempt to cross-match its tagged position with Scripps recordings was inconclusive.

Finally, the genetics. The 2024 Conservation Genetics study of 31 North Atlantic blue whales found fin whale DNA in every one, averaging 3.5 percent of the genome, flowing one way, from fin whales into blue whales. Blue-fin hybrids are breeding animals, not sterile curiosities. The 2026 Ecology and Evolution review lists 46 possible hybrids since the whaling era, 17 genetically confirmed: 15 from the North Atlantic, one from the Mediterranean, one from the northeast Pacific.

The Theories

What Doesn't Add Up

• Why has no one ever seen it? The call has been localised hundreds of times over three decades in some of the most heavily studied whale waters on Earth. Sound carries far underwater, and hydrophone fixes are coarse, but the complete absence of a visual match is still striking.
• Are the 2010 California detections the same animal or a second one? Cascadia says they are slightly lower in pitch. Nobody has published a definitive comparison.
• If CRC-BM-3330 is the singer, why did the 2015 tag-and-recording cross-match come back inconclusive? Either the whale was not calling during the tag period, the recordings did not cover it, or the hybrid is not the source.
• Why is the pitch drifting downward? Maturation, population-wide pitch decline in blue whales, or something else; the three explanations predict different futures for the signal and none has been tested against it.
• What sex is it? A male singing a blue-whale-pattern song would be unremarkable in behaviour; a female singing regularly would itself be unusual, since blue whale song is believed to be produced mainly by males.
• The Navy no longer monitors the call. Continuous tracking ended with the WHOI programme in 2004. The most detailed dataset on the animal is now more than twenty years old.

Latest Developments

The most consequential recent work is genetic. The 2024 Conservation Genetics study by Canadian and Norwegian researchers showed that North Atlantic blue whales carry an average 3.5 percent fin whale DNA and that hybrids backcross with blue whales, ending the assumption that blue-fin hybrids are sterile. In 2026, Christophe Pampoulie and colleagues published the first comprehensive review of blue-fin hybrids across ocean basins in Ecology and Evolution, cataloguing 46 possible hybrids and 17 confirmed by genetics, and a companion paper in Marine Mammal Science examined ecological competition between the two species and their hybrids. None of this identifies the 52 Hz singer, but it makes the hybrid explanation biologically routine rather than exotic.

On the public side, Netflix added The Loneliest Whale: The Search for 52 in late October 2025, prompting Cascadia Research to publish a detailed account of the 2015 expedition, the hybrid's 1995-2020 sighting record, the inconclusive 2015 cross-match, and the fact that a larger planned expedition never sailed for want of a suitable vessel. In January 2025 John Calambokidis met Joe George, the acoustician who first heard the call in 1989. No new acoustic tracking programme has been announced, and no recording of the hybrid producing the call has been published.

Sources

1. Twelve years of tracking 52-Hz whale calls from a unique source in the North Pacific (Watkins et al.) — Deep-Sea Research I, 2004
2. A Lone Voice Crying in the Watery Wilderness — Oceanus, Woods Hole Oceanographic Institution, 2005
3. 52 Hz sound - blue whale call? — NOAA Pacific Marine Environmental Laboratory
4. Maybe the World's Loneliest Whale Isn't So Isolated, After All — Smithsonian Magazine, 2015
5. The Loneliest Whale (the search for 52) recently added to Netflix: Insights into the film and Cascadia's role — Cascadia Research Collective, 2025
6. The Loneliest Whale: Documentary Chronicles a Search for the 52 Hertz Whale — Mental Floss, 2021
7. The Challenges of Filming 'The Loneliest Whale' Documentary — International Documentary Association, 2021
8. World's 'Loneliest Whale' Calls Out at a Frequency No Others Can Understand — Newsweek, 2021
9. The Loneliest Whale: The Search for 52 — Wikipedia
10. 52-hertz whale — Wikipedia
11. Hidden DNA found in blue whales reveals they've been mating with other species — and their hybrid offspring — Live Science, 2024
12. Blue-Fin Whale Hybrids Are More Common Than We Thought, Turns Out They Can Breed — IFLScience, 2024
13. What Do We Know About Hybrid Blue (Balaenoptera musculus) and Fin (B. physalus) Whales? A Comprehensive Review Across Ocean Basins (Pampoulie et al.) — Ecology and Evolution, 2026
14. Ecological Competition Between Fin Whales, Blue Whales, and Their Hybrids (Aguilar et al.) — Marine Mammal Science, 2026
15. The '52-Hertz Whale' Is The Loneliest Animal In The World. Here's What We Know — Forbes, 2025

Long Delayed Echoes

In 1927 a Norwegian engineer heard his own radio signals come back three seconds late. No one has explained why.

The Hook

A radio wave circles the Earth in about one seventh of a second. That is the limit. Once a signal has gone round the planet a few times it is absorbed, scattered, and gone. There is nothing out there, within any reasonable distance, to bounce it back.

Yet at the end of the summer of 1927, a civil engineer named Jørgen Hals sat at a receiver in Bygdøy, on the western edge of Oslo, listening to the Dutch short-wave station PCJJ in Eindhoven. He heard the normal round-the-world echo, a faint ghost trailing the telegraph signal by roughly 1/7 of a second. Then he heard something else: a second, weaker echo arriving about three seconds after the first had faded. Its strength, he estimated, was between a tenth and a twentieth of the main signal.

Three seconds is nine hundred thousand kilometres of travel. Hals could not say where it had come from. He wrote only that he could confirm he had really heard it.

Within a year the phenomenon had been observed simultaneously in two countries, published in Nature, and measured with delays of up to fifteen seconds. Then it stopped being reproducible. It has been heard, on and off, ever since.

The Facts

The documented history begins with a letter. On 29 February 1928 the Norwegian physicist Carl Størmer, already famous for measuring the height of the aurora, received Hals's account of the 1927 echoes. Størmer took it seriously enough to arrange a test. With Balthasar van der Pol of Philips in the Netherlands, and with PCJJ transmitting at a wavelength of 31.4 metres (9.55 MHz), a schedule of test signals was set up in 1928.

The result was the most compelling documentation the phenomenon has ever produced. In October 1928, echoes were recorded simultaneously in Oslo and at two locations in Eindhoven. Størmer published a short report in Nature on 3 November 1928 under the title "Short Wave Echoes and the Aurora Borealis" (Nature 122, 681). Van der Pol followed with his own note in Nature on 8 December 1928 (Nature 122, 878-879), and Edward Appleton, who would later win the Nobel Prize for his work on the ionosphere, contributed a third letter in the same issue. The Wireless World picked up the story within the month.

The delays recorded in 1928 ranged from about 3 to 15 seconds. Størmer himself was candid about the limits of his measurements: he was not adequately prepared, the times were noted by ear and by watch, and the figures gave only a qualitative idea of the effect. Van der Pol recorded that one set of observations was timed with a stopwatch and another with the second hand of an ordinary watch.

Test transmissions from PCJJ continued for about two years, at times including signals from the Dutch East Indies. Appleton's group in Britain also listened and heard echoes. In January 1930 the tests were stopped: the echoes were sporadic, the delays varied, and nobody had a mechanism to explain them. The cost could not be justified. Further experiments in France and, in 1934, by the World Radio Research League with transmitters in Britain and Switzerland produced more observations and no new understanding. Hals published his own account in the BBC's World Radio in four instalments in late 1934. He died in 1942.

Later, the definition hardened. Echoes returning more than about 2.7 seconds after transmission are classed as long delayed echoes (LDEs), the threshold chosen to exclude ordinary multi-hop and round-the-world propagation. The phenomenon has been reported from the 1.8 MHz amateur band up to 1296 MHz. It remains, by any scientific standard, an anomaly: real enough to have been recorded on instruments, rare enough that no one has ever been able to summon it on demand.

Timeline

• Late summer 1927: Jørgen Hals, at Bygdøy near Oslo, repeatedly hears a ~3-second echo of PCJJ (Eindhoven) telegraph signals.
• 29 February 1928Carl Størmer receives Hals's letter describing the echoes.
• Early 1928Test transmissions from PCJJ on 31.4 m (9.55 MHz) begin in collaboration with Balthasar van der Pol at Philips.
• 11 October 1928Størmer notes a series of echo delays, mostly around 8 seconds, ranging from 3 to 15 seconds.
• October 1928Echoes recorded simultaneously in Oslo and at two receivers in Eindhoven.
• 3 November 1928Størmer's report appears in Nature (vol. 122, p. 681).
• 8 December 1928Van der Pol and Appleton publish supporting letters in Nature (vol. 122, pp. 878-879).
• January 1930PCJJ test transmissions terminated for lack of conclusions and cost.
• 1934World Radio Research League runs echo experiments with UK and Swiss transmitters; Hals publishes his own account in World Radio.
• 1947-1949Budden and Yates transmit 27,000 vertical test signals at 13.5 and 20.7 MHz and detect no echoes (published 1952).
• May 1960Ronald Bracewell suggests in Nature that an alien probe might attract attention by repeating our signals, citing the Størmer echoes.
• 1967-1970Crawford and colleagues at Stanford record echoes with delays up to about 40 seconds at 5-12 MHz.
• 1973Duncan Lunan claims the 1928 delay sequence encodes a star map of Boötes; he later withdraws the interpretation.
• 1975Hans Rasmussen (OZ9CR) reports 4.6-second echoes at 1296 MHz in Nature.
• November 1978 to January 1979: Alan Goodacre (VE3HX) records his own 28 MHz signal returning up to about 9 seconds late.
• 1979-1980D. B. Muldrew publishes a general theory of long-delay echo generation; the "magnetospheric echo box" is explained in QST.
• 1981R. W. Freyman records several thousand echoes of up to 16 seconds at 9.9 MHz in the Alaskan auroral zone.
• 1985Vidmar and Crawford review the mechanisms and narrow the field to five likely candidates.
• 1997-2007Peter Martinez (G3PLX) logs 59 magnetospheric duct echoes of 210-220 ms over ten years in northern England.
• November 2023A Frontiers review of amateur-professional heliophysics names LDEs as an open target for HAARP-assisted citizen science.

The Evidence

The evidence for LDEs is almost entirely observational, and its quality varies enormously.

The 1928 record is the foundation. Its strength is the simultaneous reception in Norway and the Netherlands, separated by roughly a thousand kilometres, of echoes of the same PCJJ transmissions. Its weakness is the timing: delays were noted in whole seconds, by stopwatch or wristwatch, by observers who were not expecting what they heard. Every later attempt to read hidden meaning into those integers founders on this point. The numbers are approximations of a phenomenon, not measurements of a message.

The negative results matter as much. Budden and Yates transmitted 27,000 test signals over two years in the late 1940s at 13.5 and 20.7 MHz and heard nothing. This was designed to test Størmer's original hypothesis, that the echoes were reflections from electron streams far out in the Earth's magnetic field, and it effectively killed that idea.

The strongest modern data comes from the low end of the HF spectrum. Goldstone and Ellis in Hobart recorded echoes of 260-270 ms at 1.91 MHz, sometimes at a second receiver 20 km away. Martinez's decade-long British log of 59 events at 210-220 ms on 2-3.9 MHz, and observations in St Petersburg of 284-305 ms at 1.8 MHz, all fit a predictable pattern: the delay grows with the transmitter's geomagnetic latitude, the effect occurs only in darkness, and it favours winter evenings in years of low solar activity. These are now understood as magnetospherically ducted echoes. The catch is that they last a few hundred milliseconds, well below the 2.7-second threshold. They are, in Sverre Holm's phrase, medium delay echoes, and they do not explain 1928.

For the genuinely long delays the record is thinner and stranger. Goodacre's 28 MHz data from 1978-79 showed a statistical periodicity of 0.138 seconds, the round-the-world travel time, implying up to 65 circuits of the planet. Crawford's Stanford group logged delays up to 40 seconds. Freyman's Alaskan experiment produced thousands of echoes up to 16 seconds, sometimes at the same delay for transmissions minutes apart, coinciding with disturbances in the magnetic field. Rasmussen's 4.6-second echo at 1296 MHz and Yurek's chart-recorded 5.75-second echo at 432 MHz sit at frequencies that pass straight through the ionosphere, which rules out most terrestrial mechanisms.

Then there is the fraud problem. Volker Grassmann, writing in VHF Communications in 1993, warned that deliberate falsification by less serious operators could not be excluded, and recommended short transmissions on varied frequencies to expose hoaxers. A worldwide automated logging system was introduced in 2016 to reduce error and deception. The phenomenon that began with one man's word now has a filter against exactly that.

The Theories

What Doesn't Add Up

• The 1928 echoes were heard, simultaneously and at two sites, for weeks. Then a two-year professional campaign with a dedicated transmitter failed to find a pattern. What changed between the end of the summer of 1927 and January 1930?
• Budden and Yates sent 27,000 signals and heard nothing. Freyman sent signals from Alaska and heard thousands of echoes. Both were competent. Either the phenomenon depends on geography and geomagnetic conditions in ways nobody has mapped, or one of them was listening to something different.
• Delays of up to 40 seconds require the signal to travel twelve million kilometres or to be stored somewhere. There is no confirmed storage mechanism in near-Earth space that holds radio energy for that long.
• Echoes at 432 MHz and 1296 MHz pass through the ionosphere entirely. The only terrestrial explanation, Muldrew's non-linear mixing, requires an unidentified second transmitter that has never been found.
• The best-documented echoes, the magnetospheric duct events, are all under 0.5 seconds. The phenomenon that gave the field its name, echoes of many seconds, is the one with the weakest data.
• Almost a century after Hals, no observatory or space agency runs a standing LDE monitoring programme. The most persistent unexplained propagation anomaly in radio science is still investigated mainly by amateurs.

Latest Developments

The last decade has brought better tools rather than better answers. Software-defined receivers and automated logging have made it possible to timestamp echoes to the millisecond, which removes the ambiguity that undermined the 1928 data, and a worldwide automated LDE detection system announced in 2016 was designed specifically to filter out errors and hoaxes.

The most significant institutional signal came in November 2023, when a large review in Frontiers in Astronomy and Space Sciences on heliophysics and amateur radio (Frissell and colleagues) cited the study of LDEs and magnetospheric ducting as an early example of amateur-professional collaboration, and suggested that HAARP, the high-power ionospheric research transmitter in Alaska, used together with precision amateur receivers and modern modelling, could advance understanding of long delayed echoes. HAARP has since run several public campaigns with amateur participation, including lunar and asteroid echo experiments in 2022 and 2023 and an ELF/VLF campaign in early 2024, though none has yet been dedicated to LDEs.

Sverre Holm's University of Oslo reference page, the standard survey of the field, was last revised in February 2024 and still lists the same five natural mechanisms proposed in 1985. The Bygdøy echoes remain, in the terminology of the literature, unexplained.

Sources

1. Short Wave Echoes and the Aurora Borealis — Nature (C. Størmer), 1928
2. The Five Most Likely Explanations for Long Delayed Echoes — University of Oslo, Department of Physics (Sverre Holm), 2004-2024
3. 15 Possible Explanations for Long Delayed Echoes — University of Oslo (summary of Shlionskiy), 2020
4. Long delayed radio echoes: 80 years with an unexplained phenomenon — Sverre Holm (LA3ZA), lecture notes, 2009
5. Long Delayed Echoes — LA3ZA Radio and Electronics, ongoing
6. Long-delayed radio echoes: Mechanisms and observations — Journal of Geophysical Research (Vidmar and Crawford), 1985
7. Generation of long-delay echoes — Journal of Geophysical Research (D. B. Muldrew), 1979
8. Communications from Superior Galactic Communities — Nature (R. N. Bracewell), 1960
9. Long Delayed Echoes: A Study of Magnetospheric Duct Echoes 1997-2007 — RSGB RadCom (P. Martinez, G3PLX), 2007
10. Heliophysics and amateur radio: citizen science collaborations for atmospheric, ionospheric, and space physics research and operations — Frontiers in Astronomy and Space Sciences, 2023
11. Surfin': Radio Ghosts Revisited — ARRL (Stan Horzepa), 2012
12. Long Delayed Echo detection automated — Southgate Amateur Radio News (archived), 2016
13. Radio Ghosts Have Haunted the Airwaves for Nearly a Century — Vice, 2019
14. Long delayed echo — Wikipedia, accessed 2026

The Flannan Isles Lighthouse

Three keepers vanished from a locked station 110 feet above the sea. The last slate entry was 9 a.m., 15 December 1900.

The Hook

At midnight on 15 December 1900 the steamer Archtor, bound from Philadelphia to Leith, passed the Flannan Isles and saw no light. Captain Holman was certain he should have seen it. The lighthouse on Eilean Mòr, twenty miles west of Lewis, had been lit for barely a year and stood 101 metres above the Atlantic. Its beam reached twenty miles.

Nobody acted on the report. Ashore at Uig, the gamekeeper's sons who watched for the light each night had logged it missing on the 8th, 9th, 10th and 11th, seen it on the 12th, and then lost it. Haze often hid the tower for days. They did not raise the alarm.

Eleven days later the relief keeper, Joseph Moore, climbed the path to the station. The gate was shut. The doors were shut. The clock had stopped. The lamp had been cleaned and its fountain filled, ready for lighting. The kitchen was tidy, the pots washed. The beds were empty. Two sets of sea boots and oilskins were gone from their pegs. One coat, belonging to the third man, still hung there.

Three experienced keepers had left a locked station on a rock in the North Atlantic, in daylight, and had not come back.

The Facts

The lighthouse was designed by David Alan Stevenson for the Northern Lighthouse Board (NLB) and built between 1895 and 1899 by George Lawson of Rutherglen for £1,899, with a further £3,526 for the shore station at Breasclete on Lewis. Every piece of material was hauled up the 45-metre cliffs from boats. The 23-metre tower was first lit on 7 December 1899. Provisions came up from two landings, east and west, by a cable-hauled railway, with a crane at each.

In December 1900 the station was crewed by James Ducat, Principal Keeper; Thomas Marshall, Second Assistant; and Donald McArthur, an Occasional Keeper standing in for First Assistant William Ross, who was on sick leave. Joseph Moore, the Third Assistant, was ashore on rotation. Superintendent Robert Muirhead had personally recommended all three men and had made the routine relief on 7 December, when he last shook hands with them.

The station log was written up to 13 December. The entries for 14 December, and the barometer, thermometer and wind readings taken at 9 a.m. on 15 December, had been chalked on a slate for later transfer. Everything indicated that the morning's work of the 15th had been completed.

The relief tender Hesperus, under Captain Harvie, was due on 20 December but was held at Breasclete by weather. She reached the island around noon on 26 December. No flag flew and no landing preparations had been made. Harvie sounded the whistle and siren and fired a rocket. Nothing answered. Moore was put ashore at the east landing, found the station empty, and returned with the second mate McCormack to search again. Harvie left Moore, buoymaster Allan MacDonald and seamen Lamont and Campbell to keep the light and steamed back to Breasclete, from where he telegraphed the Board: "A dreadful accident has happened at Flannans."

Moore's letter of 28 December and Muirhead's report of 8 January 1901 record what was found. The east landing was untouched; its ropes lay coiled where they had been stored on 7 December. At the west landing a wooden box of mooring ropes, wedged in a crevice about 110 feet above sea level, had been torn out; its ropes were strewn among the rocks and crane legs, though every coil was still fastened. Iron railings were twisted and broken. A stone weighing more than 20 hundredweight had been carried down onto the concrete path. A lifebuoy had been ripped from its lashings. The crane, 70 feet up, was undamaged.

Muirhead concluded that on the afternoon of Saturday 15 December all three men went down toward the west landing to secure the box, and that an unexpectedly large sea rose above them and swept them away. The Crown Office decided a fatal accident inquiry was unnecessary. Ducat left a widow and four children; McArthur a widow and two. No body was ever found.

Timeline

• 7 December 1899Flannan Isles lighthouse first lit.
• 7 December 1900Routine relief made by Hesperus; Superintendent Muirhead visits and leaves the three keepers on the island.
• 8-11 December 1900: Light not seen from Uig on Lewis; seen again on the 12th.
• 13 December 1900Last written entries in the station log.
• 15 December 1900, 9 a.m.: Ducat records weather readings on the slate; morning duties completed.
• 15 December 1900, afternoon: The keepers are presumed to have gone down toward the west landing (Muirhead's reconstruction).
• 15 December 1900, midnight: SS Archtor passes the islands and sees no light.
• 18 December 1900Archtor docks at Leith.
• 20 December 1900Scheduled relief postponed by bad weather.
• 26 December 1900, noon: Hesperus arrives; Moore finds the station deserted; Harvie telegraphs the NLB that evening.
• 27-28 December 1900: Moore and the volunteers search the island end to end; Moore writes his account on the 28th.
• 28-29 December 1900: The Archtor's owners and Captain Holman report by letter and signed statement that the light was out on the 15th.
• 29 December 1900Muirhead lands with replacement keepers and begins his investigation.
• 8 January 1901Muirhead submits his report to the Commissioners.
• 1901Crown Office decides no fatal accident inquiry is required; pensions and gratuities are settled with the widows.
• 1912Wilfrid Wilson Gibson publishes the ballad "Flannan Isle", introducing the overturned chair and the untouched meal.
• 1925The station receives wireless telegraphy.
• 28 September 1971The lighthouse is automated.
• 1998Historian Mike Dash traces the "logbook entries" to later fabrication.
• 2018The feature film The Vanishing, based on the case, is released.
• December 2025Western Isles Council grants planning permission to restore the Breasclete shore station.
• June 2026Joseph Moore's letter is announced for its first public display, in Edinburgh, from 3 August 2026.

The Evidence

The primary record is small, official, and consistent: Harvie's telegram of 26 December 1900, Moore's letter of 28 December, Muirhead's report of 8 January 1901, the Archtor statements, and the minutes and correspondence of the Commissioners of Northern Lighthouses, now held by the National Records of Scotland under references NLC2/1/88, NLC2/1/89 and NLC3/1/1. The archive does not hold the station logbook; it was not among the records the Commissioners deposited.

What the record establishes is a narrow window. The slate carries readings from 9 a.m. on 15 December. The lamp had been trimmed and its oil fountains filled, which was the forenoon's work. The kitchen had been cleaned after a meal. The Archtor saw no light at midnight. The disappearance therefore falls between late morning and dusk on Saturday 15 December, and the light was not lit again until Moore lit it on the 26th.

The clothing is the sharpest physical clue. Moore told Muirhead that sea boots and oilskins were worn only when going down to the landings. Marshall's boots and oilskins were gone. Ducat's boots and an old waterproof coat were gone. McArthur's only coat remained on its peg, so he left in shirt sleeves. NLB rules required one man to remain at the light at all times. Whoever left last broke that rule.

The west landing damage is the strongest evidence of violence. Muirhead wrote that some of it was difficult to believe unless actually seen. The rope box at 110 feet was gone. Railings at that height were bent flat. A stone of more than a ton had moved. A lifebuoy had been torn from ropes that were themselves untouched, with canvas still clinging to the lashings. R. W. Munro, drawing on the same records, adds that turf had been stripped from the clifftop more than 200 feet above the sea.

What is not evidence should be stated as plainly. There was no overturned chair and no half-eaten meal; both come from Gibson's 1912 poem, and Moore's own account says the kitchen utensils were clean. There were no logbook entries describing a keeper weeping, the men praying, or "God is over all". The log ended on 13 December, later entries were on a slate, and the quoted lines have been traced by Mike Dash to invention in later popular accounts, cross-cited until they acquired false authority.

The remaining facts are absences. No bodies. No witnesses. No gate left open. Four men walked the island from end to end within two days and found nothing.

The Theories

What Doesn't Add Up

• Why was the light out on the 8th, 9th, 10th and 11th of December, four nights before the men vanished? Muirhead accepted that haze often hid the tower for four or five nights, but it is an uncomfortable coincidence in a reconstruction that depends on everything being normal until the 15th.
• Why did McArthur leave without his coat, in December, in weather violent enough to lift a ton of rock? Either he ran in an emergency or he never expected to be outside for long.
• Who shut the gate and the doors? Moore found both closed. A man sprinting to warn colleagues does not usually latch a gate behind him.
• The rope box was 110 feet above sea level. Muirhead's theory requires a wave to rise above that height and come down on the men from above. Storm-driven seas of that scale exist, but the theory has never been tested against wave records for 15 December 1900.
• All three left the light unattended, against standing orders, in daylight, at the same time. The most disciplined explanation is a chain of rescue attempts, but that is inference, not evidence.
• The station logbook, the one document that would show the keepers' own words for 13 December, was never deposited with the national archive and its whereabouts are unknown.
• The Crown Office declined to hold a fatal accident inquiry. The only investigation was internal, conducted by the man who had recruited the missing keepers.

Latest Developments

The case has entered a second life as heritage. The Northern Lighthouse Board now publishes the Harvie telegram, the Moore letter and the full Muirhead report on its website, and has noted the 2018 film The Vanishing (directed by Kristoffer Nyholm, with Gerard Butler, Peter Mullan and Connor Swindells as the keepers) while reminding readers that three men died and their families never learned their fate. The film invents a chest of gold and a spiral of violence; it is drama, not reconstruction. Emma Stonex's 2021 novel The Lamplighters reworked the disappearance again.

In December 2023 the National Records of Scotland published a blog account drawn from the Commissioners' files, and its lighthouse research guide, updated in February 2026, lists every surviving document. In December 2025 Western Isles Council granted planning permission to restore the Breasclete shore station as exhibition space and holiday flats.

The most direct development came in June 2026: Joseph Moore's handwritten letter of 28 December 1900, reference NLC3/1/1, was announced for its first public display in the free exhibition "For the Safety of All: Scotland's Lighthouses" at General Register House, Edinburgh, from 3 August to 25 September 2026. It is the closest thing to a witness statement the case will ever have.

Sources

1. Flannan Isles — Northern Lighthouse Board (Harvie telegram, Moore letter, Muirhead report of 8 January 1901), accessed 2026
2. The mystery of the Flannan Islands Lighthouse — National Records of Scotland research guide, 2026
3. Flannan Isles Lighthouse Keepers: The disappearance — National Records of Scotland blog, 2023
4. Letter reporting lighthouse keepers' mysterious 1900 disappearance to go on display for the first time — Who Do You Think You Are? Magazine, 2026
5. New Edinburgh Fringe exhibition to document the live-saving stories of Scotland's lighthouses — The Scotsman, 2026
6. Letter on mystery of three missing lighthouse keepers to go on display — BBC News, 2026
7. True Strange Stories? — Fortean Times (Mike Dash), 2017
8. Skeptoid #610: The Keepers of Flannan Light — Skeptoid (Brian Dunning), 2018
9. Flannan Isle Mystery: When Three Lighthouse Keepers Vanished Forever — History Hit, 2021
10. The Flannan Isle mystery: The three lighthouse keepers who vanished — Sky HISTORY, 2020
11. The Vanishing (2018 film)) — Wikipedia, accessed 2026
12. The Lamplighters by Emma Stonex review — The Guardian, 2021
13. Flannan Isles: the lighthouse keepers who vanished — Boing Boing, 2026
14. Flannan Isles Lighthouse — Wikipedia, accessed 2026

The Kecksburg Incident

A fireball crossed six states, the Army sealed a Pennsylvania wood, and NASA later admitted the files were lost.

The Hook

At 4:43 p.m. Eastern time on Thursday, 9 December 1965, a seismograph 25 miles southwest of Detroit recorded a shock wave. The Federal Aviation Administration logged 23 reports from pilots. Across southern Ontario and at least six American states, people saw a brilliant fireball crossing the sky in daylight. Hot metal fragments were reported in Michigan and Ohio. Grass fires started near Elyria and Lapeer. Sonic booms rattled the Pittsburgh area.

Three minutes later, about thirty miles southeast of Pittsburgh, residents of a village called Kecksburg reported blue smoke rising from a wood, vibrations, and a thump.

That evening the Greensburg Tribune-Review reported that the area had been sealed off on the orders of United States Army and State Police officials, pending a close inspection of whatever had fallen. The next day the same paper ran a different headline: "Searchers Fail to Find Object". The Air Force called it a meteor.

Forty years later, a NASA spokesman told the Associated Press that the agency had examined fragments from the site, identified them as Soviet, and lost the records. NASA's own chief orbital debris scientist had already said no satellite could have been there. Both statements are on the record. They cannot both be true.

The Facts

The fireball is not in dispute. Sky & Telescope, in February 1966, placed it over the Detroit-Windsor area at about 4:44 p.m. EST on a path running roughly northwest to southeast, ending in or near western Lake Erie. In 1967 astronomers Von Del Chamberlain and David J. Krause published a detailed analysis in the Journal of the Royal Astronomical Society of Canada. Using the seismic record to fix the time at 4:43 p.m., and two photographs of the smoke trail taken north of Detroit to triangulate the trajectory, they found the object descending steeply and concluded it probably fell near the northwestern shore of Lake Erie near Windsor. Astronomers Paul Annear, William Bidelman and Fred Hess, and geophysicist George Wetherill, all told the press it was a meteor.

What happened at Kecksburg is contested, but some of it is documented. On the evening of 9 December, state police, volunteer firefighters and Air Force personnel searched the woods. Reporters, including Bob Gatty of the Tribune-Review and John Murphy of radio station WHJB, were on the scene and later described a military presence. By the early hours of 10 December the search had officially found nothing; an AP report said searchers had tramped through the area for hours with Geiger counters.

One official document survives. A Project Blue Book memorandum dated 10 December 1965, uncovered in the mid-1980s by researcher Ray Boeche, records that a three-man team from the 662nd Radar Squadron at Oakdale, Pennsylvania, was dispatched to "Acme" to investigate and pick up an object that had started a fire. It also records a call from a Major Howard at the Pentagon to Major Hector Quintanilla, head of Blue Book, asking what could be told to the public. Quintanilla said the team had searched unsuccessfully, that it was "Ok to call it a meteor", that the investigation was still under way, and that no space debris had entered the atmosphere that day. Witnesses near the site had Acme mailing addresses, which is why the file was never indexed under Kecksburg.

The witness testimony, gathered over decades chiefly by Greensburg researcher Stan Gordon, describes an acorn-shaped metallic object roughly the size of a Volkswagen Beetle, bronze or gold in colour, seamless, with a band of raised markings resembling hieroglyphs around its base, semi-buried in the ground. Witnesses including Bill Bulebush, Jim Romansky and John Hays described it being removed on a military flatbed truck under a tarpaulin. None of this was reported in 1965; it emerged after the NBC series Unsolved Mysteries revived the case in 1990.

Kosmos 96, a Soviet Venus probe stranded in low orbit after a launch failure on 23 November 1965, re-entered on 9 December 1965. It is the most frequently proposed man-made candidate, and the one most decisively excluded.

Timeline

• 23 November 1965Kosmos 96 launched from Baikonur; a third-stage failure leaves it in a 227 x 310 km orbit.
• 9 December 1965, 3:18 a.m. EST: US Space Command data later places the re-entry of Kosmos 96 over Canada.
• 9 December 1965, 4:43-4:44 p.m. EST: Fireball seen over Ontario, Michigan, Ohio, Indiana, Pennsylvania, West Virginia and New York; seismograph and 23 pilot reports record it.
• 9 December 1965, about 4:46 p.m.: Kecksburg residents report smoke, vibration and a thump in the woods.
• 9 December 1965, evening: State police, firefighters, Air Force personnel and reporters converge on the site; the Tribune-Review reports the area sealed by Army and State Police.
• 10 December 1965Official search ends with nothing found; Air Force declares a meteor; Blue Book memo records the Oakdale team and the "Ok to call it a meteor" exchange.
• February 1966Sky & Telescope publishes its "Great Lakes Fireball" analysis.
• 1967Chamberlain and Krause triangulate the trajectory in the JRASC and place the terminal point near Windsor, Ontario.
• Mid-1980s: Ray Boeche locates the Blue Book "Acme" file.
• 1987NASA's "fragology" files, covering recovered space objects 1962-1967, recorded as missing from the Federal Records Center.
• 1990Unsolved Mysteries airs the Kecksburg segment; witnesses describe the acorn object on national television.
• 1991James Oberg proposes Kosmos 96 as the explanation.
• 2002The Sci Fi Channel sponsors a Freedom of Information initiative led by journalist Leslie Kean.
• 2003Kean sues NASA under FOIA in federal court in Washington; NASA chief orbital debris scientist Nicholas Johnson states categorically that Kosmos 96 could not have reached Pennsylvania at 4:45 p.m.
• 24 October 2003The Sci Fi Channel broadcasts "The New Roswell: Kecksburg Exposed", hosted by Bryant Gumbel.
• December 2005NASA spokesman David Steitz tells AP the fragments were from a Soviet satellite and the records were lost.
• October 2007Settlement before Judge Emmet Sullivan; NASA agrees to a court-monitored search and pays legal fees; public liaison Steve McConnell testifies that two boxes are missing.
• June-July 2008: NASA reports 297 boxes searched and releases several hundred pages, none mentioning Kecksburg.
• June-August 2009: Final search completed; joint motion for dismissal filed; Kean's report published in November.
• December 201550th anniversary; John Ventre and Owen Eichler propose a GE Mark 2 re-entry vehicle; witness Romansky rejects it.
• 9 December 202560th anniversary; WTAE-TV locates two reels of 1965 archival film.
• 12 June 2026Steven Spielberg's Disclosure Day, which references Kecksburg, is released.

The Evidence

The physical evidence for the fireball is good and points away from Kecksburg. Seismic timing, pilot reports, two photographs of the smoke trail and the fires and metal fragments reported in Michigan and Ohio are consistent with a large bolide on a steep, prograde trajectory that ended over western Lake Erie. The 1967 JRASC analysis, the only peer-reviewed reconstruction, puts the terminal point roughly 200 miles from Kecksburg. NASA's own catalogue entry for Kosmos 96 states that the fireball's path was probably too steep for a spacecraft re-entering from orbit and more likely a meteor from the asteroid belt.

The evidence against Kosmos 96 is stronger still. US Air Force tracking data indicate its orbit decayed before 21:43 UTC on 9 December. US Space Command's reconstruction places the re-entry over Canada at 3:18 a.m., thirteen hours before the fireball. Nicholas Johnson, NASA's chief scientist for orbital debris from 1996 to 2014, told Kean in 2003 that there was no way any Kosmos 96 debris could have landed in Pennsylvania around 4:45 p.m., and that no other known man-made object re-entered that day. Data from US Space Command and the Russian Space Agency agree.

The evidence for something at Kecksburg is documentary and testimonial. The Tribune-Review's same-day report of an Army and State Police cordon was written by a reporter on the scene; it is contemporaneous and specific. The Blue Book memo confirms that a military team was sent to pick up an object, and that the "meteor" explanation was authorised while the investigation was still open. A 1969 NASA memo released in the lawsuit, written by fragology custodian Richard Schulherr, records that when a re-entry could not be identified, NASA's practice was to call it a meteorite or an untracked satellite.

The witness accounts are the weakest link and the most vivid. None of the acorn descriptions were recorded in 1965; they surfaced 25 years later and have grown in detail since. In 2012 Joe Nickell and James McGaha in Skeptical Inquirer called this the "Roswellian syndrome", in which an ordinary event acquires a crashed-craft narrative through retrospective testimony. Robert R. Young's 2023 study "Metamorphosis" traces how the accounts changed over time.

The evidence from the lawsuit is a list of absences. NASA initially said it had no responsive documents; hundreds later appeared. More than 300 boxes were searched. The two boxes of fragology files (accession 68A2062), described as reports of space object recoveries and analysis of fragments to determine national ownership, were confirmed missing since 1987. Seven boxes of NASA-DOD agreements for June 1965 to May 1966 were checked out by a NASA records manager on 4 December 1996 and never returned. Box 7, NASA-DOD 1964-66, was destroyed. Twenty requested boxes could not be produced. Not one page mentioned Kecksburg.

The Theories

What Doesn't Add Up

• The Tribune-Review reported an Army and State Police cordon on the night of 9 December, from a reporter on the ground. Within a day the story became "nothing found". Both cannot be complete.
• The Blue Book memo says a team was sent to pick up an object that started a fire, and that it was acceptable to call it a meteor while the investigation continued. Why authorise an explanation before finishing the inquiry?
• NASA's spokesman said in 2005 that fragments had been examined and identified as Soviet. NASA's chief orbital debris scientist said in 2003 that no Soviet object could have been there. The agency has never reconciled the two statements.
• The fragology files, the one NASA record series explicitly about recovered space objects and their national origin for 1962-1967, went missing in 1987. Seven boxes of NASA-DOD agreements covering the exact period were checked out in 1996 and never returned.
• NASA's search produced clippings on lesser fireballs from the same period, but none on a fireball seen across six states and reported in the New York Times.
• The report written by the 662nd Radar Squadron's team after its search has never been found, although its author supplied an affidavit that it existed.
• No witness described an acorn in 1965. The first public accounts came in 1990. Either the story grew, or people were silent for a quarter century.
• The Kecksburg fire department's incident logs for 9 December 1965 are missing; the rest of the month's records survive.

Latest Developments

The sixtieth anniversary in December 2025 produced the most activity in years. The village's annual UFO Festival, held on the fire department grounds in July 2025, marked its twentieth edition. In November 2025 Stan Gordon published an extended report setting out, for the first time, the more extreme second-hand claims he had collected: tracks leading from the object, a hatch, a stretcher, and non-human remains at Wright-Patterson, plus a 2024 account of an F-101 fighter firing a missile at the object near Pontiac, Michigan. None of these claims is corroborated by any document, and the witnesses are almost all deceased.

More concrete was the announcement on 4 December 2025 that WTAE-TV anchor Michelle Wright, working with the station's archive manager, had located two reels of 1965 film footage related to the incident, broadcast in two-part reports on 7 and 14 December 2025. Gordon had spent years asking Pittsburgh stations for such footage and been told none survived.

Steven Spielberg's Disclosure Day, released on 12 June 2026, presents Kecksburg within its fiction as one of a series of released government recordings, prompting fresh coverage in Pennsylvania media. No new official document has been released since NASA's search closed in 2009.

Sources

1. The Conclusion of the NASA Lawsuit: Concerning the Kecksburg, PA UFO case of 1965 — Leslie Kean / Coalition for Freedom of Information, 2009
2. Forty Years of Secrecy: NASA, the Military, and the 1965 Kecksburg Crash — International UFO Reporter (Leslie Kean), 2005
3. Is case finally closed on 1965 UFO mystery? — NBC News / Space.com (Leonard David), 2009
4. Judge forces NASA to take a giant leap in FOIA suit — Reporters Committee for Freedom of the Press, 2007
5. Nasa told to solve 'UFO crash' X-File — The Guardian, 2007
6. Cosmos 96, NSSDCA/COSPAR ID 1965-094A — NASA Space Science Data Coordinated Archive, 2018
7. Greensburg Tribune-Review headline story, December 10, 1965 — Greensburg Tribune-Review (archived), 1965
8. Jim Oberg on the "NASA lawsuit over Kecksburg UFO Documents" — James Oberg, 2008
9. The Roswellian Syndrome: How Some UFO Myths Develop — Skeptical Inquirer (Nickell and McGaha), 2012
10. Five decades later, the Kecksburg UFO is identified (probably) — Pittsburgh Post-Gazette, 2015
11. December 9, 2025, Will Mark The 60th Anniversary of the Kecksburg, PA UFO Incident — Stan Gordon's UFO Anomalies Zone, 2025
12. WTAE-TV News Anchor Locates Archival Film Footage from 1965 — Stan Gordon's UFO Anomalies Zone, 2025
13. 20th UFO Festival celebrates 60 years of mystery in Kecksburg — Latrobe Bulletin, 2025
14. Long rumored alien encounter in 'Pennsylvania's Roswell' is referenced in Steven Spielberg's 'Disclosure Day' — PhillyVoice, 2026
15. Kecksburg UFO incident — Wikipedia, accessed 2026

The Dancing Plague of 1518

One woman began dancing in a Strasbourg street on 14 July 1518. Within a month, chronicles counted 400.

The Hook

The Strasbourg city archives still hold the order. In the summer of 1518, the council of one of the largest cities in the Holy Roman Empire banned music and public dancing, with a single exception written into the record by the chancellor Sebastian Brant, author of The Ship of Fools: honourable persons could still dance at weddings, but only to stringed instruments, and on their conscience not to use tambourines or drums.

A city government does not legislate against drums without cause. The cause, according to municipal notes, physicians' reports, cathedral sermons and six separate chronicle accounts, was that a woman had stepped into the street on or about 14 July and started to dance. She did not stop. Within days, dozens had joined her. Within a month, one family chronicle put the number at 400. Their feet bled into their shoes. Some, later writers claimed, collapsed and died.

The council's first remedy was to build stages and hire musicians. It made things worse. What follows is what the documents actually say, what they do not say, and why the best available explanation still leaves a hole at the centre of the case.

The Facts

The outbreak is unusually well documented for a sixteenth-century event. Historians including John Waller (Michigan State University), H. C. Erik Midelfort (University of Virginia) and Lynneth Miller Renberg (Anderson University) have worked from Strasbourg council minutes, the chronicles of Sebastian Brant, Daniel Specklin and the Imlin family, sermons, and the writings of the physician Paracelsus, who visited the city in 1526 and later described the episode in his Opus Paramirum.

The sources agree on the core sequence. In mid-July 1518 a single woman began dancing in a Strasbourg street. Four of the six chronicle accounts name her as Frau Troffea; the others leave her anonymous. She danced for days, reportedly four to six, before the authorities sent her by wagon to the shrine of St Vitus at Saverne, roughly thirty miles away. By then others had begun to imitate her. Within a week the number of dancers had passed thirty. The Imlin chronicle records 400 afflicted within a month. Modern estimates range from 50 to 400, and Waller accepts the higher figure as plausible.

The city's response is also on the record. The council consulted the physicians' guild, which declared the dancing "a natural disease, which comes from overheated blood" and, rather than bleeding the patients, prescribed that they dance the affliction out of their bodies. Specklin's chronicle describes the council ordering the carpenters' and tanners' guild halls opened as dance floors and platforms set up in the horse market and the grain market. Musicians were paid to play; healthy dancers were hired to keep the sick moving.

When the numbers grew rather than fell, the council reversed course. It ordered the stages torn down, banned public dancing and most music, and had the worst cases loaded onto wagons for the three-day journey to Saverne. There, priests placed the dancers under a carved image of St Vitus, put small crosses in their hands and red shoes on their feet, and sprinkled the shoes with holy water and consecrated oil. Word came back that the pilgrims were cured. By late August or early September, the outbreak had subsided.

This was not an isolated event. Waller notes that the 1374 dancing epidemic that began around Aachen and spread along the Rhine is attested by dozens of chronicles, and that more than ten such outbreaks occurred along the Rhine and Moselle between 1374 and 1518. Strasbourg's was the last major one and the best recorded. By the mid-1600s, compulsive-dancing outbreaks had disappeared from Europe.

Timeline

• 1374 – A dancing epidemic breaks out around Aachen and spreads along the Rhine; the first well-attested outbreak of "St John's" or "St Vitus'" dance.
• 1492 – A comet over Alsace feeds apocalyptic expectation; the decades that follow bring floods, hard frosts, plague and famine to the region.
• 1517 – A third consecutive year of failed harvests around Strasbourg; peasant unrest and widespread malnutrition are recorded.
• 14 July 1518 – Frau Troffea (as four of six chronicles name her) begins dancing in a Strasbourg street and does not stop.
• c. 18–20 July 1518 – After four to six days, the authorities send Troffea by wagon to the shrine of St Vitus at Saverne; more than thirty others are already dancing.
• Late July 1518 – The physicians' guild diagnoses "overheated blood"; the council opens guild halls and builds stages in the horse and grain markets, hiring musicians.
• August 1518 – The Imlin family chronicle records 400 dancers within a month of the first case.
• August 1518 – The council dismantles the stages and bans dancing and music; Sebastian Brant records the exception for weddings with stringed instruments only.
• August 1518 – Wagons of dancers are sent to Saverne, where priests perform the red-shoe ritual under the image of St Vitus.
• Late August / early September 1518 – The outbreak subsides; the stream of pilgrims to Saverne dwindles to a trickle.
• 1526 – Paracelsus visits Strasbourg, gathers accounts and later writes about the episode in his Opus Paramirum, blaming Troffea's "lewd" imagination and classifying the illness as chorea lasciva.
• 1636 – An astrological chronicle by Goldmeyer describes dancers who "fell down unconscious and succumbed to death", the source of later death claims.
• 1832 – German physician J. F. C. Hecker publishes Die Tanzwuth, framing the medieval dancing manias as a single epidemic disease.
• 1952 – Swedish pharmacologist E. Louis Backman argues for ergot poisoning in Religious Dances in the Christian Church and in Popular Medicine.
• 1999 – H. C. Erik Midelfort's A History of Madness in Sixteenth-Century Germany re-examines the Strasbourg archival record.
• 2008 – John Waller publishes A Time to Dance, a Time to Die, arguing for a stress-induced mass psychogenic illness; a Lancet summary follows in February 2009.
• 2016 – Élisabeth Clementz, reviewing the French edition of Waller's book in Revue d'Alsace, notes that the city's contemporary sources record no deaths.
• 2017 – Lynneth Miller's Dance Research paper compares the six chronicle accounts and the church's and physicians' rival diagnoses.
• 2020 – Jonathan Glazer's short film Strasbourg 1518, made in lockdown for the BBC and Artangel, brings the case back to public attention.

The Evidence

The evidence is entirely documentary. There are no bodies, no burial registers linked to the outbreak, and no physical traces. What survives is a layered paper record, and the layers do not all carry the same weight.

The strongest layer is contemporary and administrative. The Strasbourg municipal archives contain the council's own notes on the crisis, including the music-and-dancing ban (Midelfort cites the ordinance at R3, fol. 72 recto). Sebastian Brant, the city chancellor, was an eyewitness official, and his annals preserve the physicians' "overheated blood" verdict. These sources establish beyond reasonable doubt that a large number of people danced compulsively for weeks and that the city treated it as an emergency.

The second layer is near-contemporary narrative. Daniel Specklin, the city architect, wrote his chronicle later in the sixteenth century and supplies the detail about guild halls, platforms and the Saverne ritual. The Imlin family chronicle supplies the figure of 400. Paracelsus, writing in the late 1520s or 1530s after a 1526 visit, is the earliest source to name Frau Troffea and the earliest to attempt a psychological explanation. He is also, as Ned Pennant-Rea observes, a hostile witness: he believed Troffea was faking to embarrass her husband.

The third layer is late and unreliable. The claim that fifteen people died each day at the peak, repeated in most popular accounts and in Waller's book, derives from chronicles compiled a century or more after the event, such as the 1636 Goldmeyer chronicle and Johann Schilter's seventeenth-century compilation quoting a lost poem. Wikipedia's editors, citing Clementz's 2016 review, note that the Strasbourg sources of 1518 do not record any deaths, or even state that there were fatalities. That is the single most important caveat in this file: the "deadly" dancing plague is deadly only in the retellings.

What the documents do agree on is the clinical picture. Victims moved spasmodically, arms flailing, eyes described as glassy or vacant, bodies drenched in sweat, feet swollen and bleeding. They appeared unable to stop, cried out for help, and collapsed from exhaustion before rising to dance again. The numbers, though contested, are consistent in scale across independent sources: thirty-odd within a week, hundreds within a month.

The contextual evidence is also solid. Regional chronicles record repeated harvest failures in 1515–1517, floods, a bitter winter, the arrival of syphilis in the 1490s, recurring plague and leprosy, and peasant unrest. Strasbourg's poor in the summer of 1518 were malnourished, frightened and, in Waller's phrase, primed.

The Theories

What Doesn't Add Up

• Did anyone die? The most repeated fact about the case, fifteen deaths a day, rests on chronicles written more than a century later. The city's own 1518 records mention no fatalities. Either the deaths were real and the council chose not to record them, or the deadliest detail of the dancing plague was added by later writers.
• Why 1518, and why only Strasbourg? Famine, plague and religious dread were general across the Rhineland. No other city recorded an outbreak that summer, and no explanation yet identifies what made Strasbourg the only place where the script played out.
• How did people physically sustain it? Continuous dancing for days, on injured feet, with little food or water, exceeds what documented modern episodes of mass psychogenic illness produce. Either the chroniclers exaggerated duration, or the trance state was deeper than current models allow.
• Who was Frau Troffea? Four chronicles name her; two do not. Nothing else is known: no age, no parish, no husband's name, no record of what became of her after Saverne.
• Why did the outbreaks stop? Dancing manias recur along the Rhine and Moselle for roughly 150 years and then vanish by the mid-1600s. Waller links this to the decline of belief in St Vitus' curse, but the correlation has never been tested against the archival record of other towns.

Latest Developments

There has been no major archival discovery since the 500th anniversary in 2018, but scholarly attention has shifted. A Dance Research study by Lynneth Miller (2017) compared the six chronicle accounts in detail and framed the episode as a contest between two authorities, the church treating it as divine punishment and the physicians as a natural disease. National Geographic's October 2025 feature, quoting Miller Renberg and the medieval dance historian Kathryn Dickason, reflected the current mood: mass psychogenic illness remains the leading explanation, but historians increasingly argue that asking only "what caused it" misses what the outbreak meant to the people who recorded it, and that gender, class and the coming witch-hunts shaped who became a victim. Clementz's 2016 observation that contemporary sources record no deaths has percolated into reference works, and the "fifteen a day" figure is now routinely flagged as unverified. In popular culture, Jonathan Glazer's 2020 film Strasbourg 1518 and Florence + the Machine's 2022 track "Choreomania" have kept the case in circulation, without adding evidence.

Sources

1. A forgotten plague: making sense of dancing mania60386-X/fulltext) — The Lancet (John Waller), 2009
2. A forgotten plague: making sense of dancing mania (PubMed record) — PubMed / The Lancet, 2009
3. The Dancing Plague of 1518 — The Public Domain Review (Ned Pennant-Rea), 2018
4. Divine Punishment or Disease? Medieval and Early Modern Approaches to the 1518 Strasbourg Dancing Plague — Dance Research, Edinburgh University Press (Lynneth Miller), 2017
5. What caused Strasbourg's dancing plague of 1518? — National Geographic (Olivia Campbell), 2025
6. Dancing plague of 1518 — Wikipedia, accessed 2026
7. Dancing plague of 1518: Facts & Theories — Encyclopaedia Britannica, accessed 2026
8. What Was the Dancing Plague of 1518? — History.com (Evan Andrews), 2015, updated 2025
9. A Time to Dance, a Time to Die: The Extraordinary Story of the Dancing Plague of 1518 (review) — German History, Oxford Academic, 2009
10. A History of Madness in Sixteenth-Century Germany — Stanford University Press (H. C. Erik Midelfort), 1999
11. Waller (John), Les danseurs fous de Strasbourg (review) — Revue d'Alsace (Élisabeth Clementz), 2016
12. The black death and the dancing mania — J. F. C. Hecker, 1888 English edition, Internet Archive
13. Strasbourg 1518: reliving a 16th-century 'dancing plague' in lockdown — The Guardian, 2020
14. Dancing death — BBC News (John Waller), 2008

Fast Radio Bursts

Millisecond radio flashes from other galaxies, thousands a day. One was traced to a magnetar. The rest are open.

The Hook

The signal had been sitting on a disk for six years. In 2007, Duncan Lorimer and his student David Narkevic at West Virginia University were combing through archived data from the 64-metre Parkes radio telescope in Australia when they found a pulse recorded on 24 July 2001. It lasted about five milliseconds. It was, for that instant, brighter than any other radio source in that patch of sky. And it was smeared across frequencies in a way that told them how much ionised gas it had crossed: a dispersion measure of roughly 375 parsecs per cubic centimetre, several times more than the Milky Way could account for in that direction.

The only reasonable reading was that the flash had come from far outside our galaxy. Nothing known could put that much energy into radio waves in a few thousandths of a second and then go silent.

For six years afterwards it was a single event, and some astronomers suspected an instrumental glitch. Then more turned up. Today the count is in the thousands, a handful have been pinned to individual galaxies and one to a single star. The question that remains is the one Lorimer's team asked first: what is doing this?

The Facts

A fast radio burst (FRB) is a transient pulse of radio emission lasting from a fraction of a millisecond to a few seconds, with a total energy that, in the brightest cases, matches what the Sun radiates over decades. The defining measurable property is dispersion: lower frequencies arrive later than higher ones because the signal has passed through free electrons in space. The delay gives the dispersion measure (DM), and for almost all FRBs the DM far exceeds the contribution of our own galaxy. That excess is the evidence they are extragalactic.

The first, now called FRB 010724 or the Lorimer burst, was reported in Science in 2007. It was followed in 2013 by four more at high Galactic latitude found by Dan Thornton and colleagues in the Parkes High Time Resolution Universe survey, which established that FRBs were a population at cosmological distances, not a fluke. A confusion did follow: Parkes had also been recording "perytons", dispersed pulses that turned out, in a 2015 investigation led by Emily Petroff, to come from the observatory's own microwave ovens being opened mid-cycle. The same study showed the ovens could not explain the Lorimer burst.

The second decisive fact came from Arecibo. A burst recorded on 2 November 2012, FRB 121102, was found in 2016 to repeat. Repetition ruled out a one-time cataclysm for that source and let astronomers point an interferometer at it. In 2017 Shami Chatterjee and colleagues used the Very Large Array to localise it to a faint, low-metallicity dwarf galaxy at redshift 0.193, roughly three billion light-years away, sitting beside a persistent radio source.

Since 2018 the Canadian Hydrogen Intensity Mapping Experiment (CHIME) in British Columbia, with a 200-square-degree field of view, has detected FRBs at a rate approaching two per day. Its first catalogue (2021) contained 536 bursts, including 62 from 18 repeating sources. Its second catalogue, released in 2025 and highlighted by Nature Astronomy in March 2026, holds 4,539 bursts from 3,641 sources, 83 of them repeaters. Repeaters and apparent one-offs are spread uniformly across the sky but differ in burst shape and bandwidth.

Two further facts anchor the case. On 28 April 2020, CHIME and the small STARE2 array in California independently detected a millisecond radio burst from SGR 1935+2154, a known magnetar in the Milky Way, coincident with an X-ray burst. It was roughly a thousand times more energetic than any previous magnetar radio burst and only about ten times weaker than the faintest extragalactic FRBs. And in 2020 Jean-Pierre Macquart and colleagues used a handful of FRBs localised by Australia's ASKAP telescope to show that the DM–redshift relation accounts for the universe's "missing" ordinary matter, which resides as diffuse ionised gas between galaxies. FRBs had become a tool as well as a mystery.

Timeline

• 24 July 2001 – Parkes records a five-millisecond dispersed pulse near the Small Magellanic Cloud; nobody notices for six years.
• 2007 – Lorimer, Bailes, McLaughlin, Narkevic and Crawford publish the burst in Science; DM about 375 pc cm⁻³ implies an extragalactic origin.
• 2 November 2012 – Arecibo detects FRB 121102 (published 2014), the first FRB found outside Parkes.
• 2013 – Thornton et al. report four FRBs from the Parkes HTRU survey; the name "fast radio burst" enters use.
• 2015 – Petroff et al. trace Parkes "perytons" to on-site microwave ovens, clearing FRBs of the same suspicion.
• 2016 – Spitler et al. announce that FRB 121102 repeats.
• 2017 – Chatterjee et al. localise FRB 121102 with the VLA to a dwarf galaxy at z = 0.193; Marcote et al. refine the position to milliarcseconds with the European VLBI Network.
• July 2018 – CHIME/FRB begins operations and starts detecting bursts daily.
• 2019 – ASKAP's CRAFT team localises a non-repeating FRB to a massive galaxy, proving one-offs can be pinned down too.
• 2020 – CHIME reports a 16.35-day activity cycle in the repeater FRB 20180916B, the first periodic FRB.
• 28 April 2020 – CHIME and STARE2 detect FRB 200428 from the Galactic magnetar SGR 1935+2154.
• May 2020 – Macquart et al. publish the FRB census of baryons in Nature, resolving the missing-baryon problem.
• June 2021 – The first CHIME/FRB catalogue: 536 bursts, 18 repeaters.
• February 2022 – Kirsten et al. localise FRB 20200120E to a globular cluster in M81, about 12 million light-years away, an old stellar environment.
• October 2023 – Ryder et al. report FRB 20220610A at redshift about 1, whose light travelled roughly eight billion years.
• January 2025 – Two papers place the repeater FRB 20240209A about 130,000 light-years outside a massive, quiescent elliptical galaxy 11.3 billion years old.
• 16 March 2025 – CHIME and its three new Outrigger stations detect FRB 20250316A, "RBFLOAT", and localise it to 13 parsecs in NGC 4141, 130 million light-years away.
• 2025 – The second CHIME/FRB catalogue: 4,539 bursts from 3,641 sources, 83 repeaters.
• January 2026 – A Science paper on FRB 220529A reports a hundredfold "rotation-measure flare" consistent with plasma ejected by a companion star, evidence for a binary system.
• March 2026 – Publication of JWST near-infrared observations at the RBFLOAT site, showing a faint counterpart and no repeat bursts in more than six years of CHIME data.

The Evidence

Everything known about FRBs comes from the pulses themselves and, increasingly, from where they land on the sky.

Duration and energy. Bursts last from under a millisecond to a few seconds. Light-travel-time arguments limit the emitting region to at most a few hundred kilometres for the shortest bursts, the size of a neutron star or a compact region near one. The energies are extreme: FRB 20220610A, at redshift about one, released in a fraction of a second the equivalent of the Sun's output over thirty years.

Dispersion and scattering. The Lorimer burst's DM of about 375 pc cm⁻³ was the first clue; catalogued DMs now run from under 100 to several thousand. The Macquart relation ties DM to redshift with enough reliability that seven localised FRBs sufficed in 2020 to measure the cosmic baryon density. Scattering tails and scintillation in the pulses probe turbulent gas in host galaxies and along the line of sight.

Polarisation. Many FRBs are close to 100 per cent linearly polarised. Faraday rotation gives a rotation measure (RM) that tracks the magnetic environment. FRB 121102 sits in an extremely magnetised region; FRB 220529A's RM jumped by more than a factor of a hundred in late 2023 and relaxed within two weeks, which the discovery team interprets as a coronal mass ejection from a companion crossing the line of sight.

Repetition and periodicity. About three per cent of known sources repeat, though the true fraction is debated because a "one-off" may simply not have been observed long enough. FRB 20180916B is active in a roughly five-day window every 16.35 days, with the window arriving earlier at higher frequencies. Repeater bursts tend to be narrower in bandwidth and longer in duration than one-offs, a statistical difference confirmed across CHIME's 83 repeaters.

Localisation and hosts. The count of arcsecond-localised FRBs stood at about two dozen in 2022 and has reached nearly 200 by 2026, thanks to ASKAP, the Deep Synoptic Array (DSA-110, which has found about 60 FRBs and localised about 30), MeerKAT and the CHIME Outriggers in British Columbia, West Virginia and California. Hosts include star-forming dwarfs (FRB 121102), a spiral arm (FRB 20180916B), a globular cluster (FRB 20200120E), the far outskirts of a dead elliptical galaxy (FRB 20240209A), and a nearby spiral at 130 million light-years (RBFLOAT). No single stellar population fits them all.

The Galactic anchor. FRB 200428 from SGR 1935+2154 remains the only FRB-like burst tied to an identified object, with a STARE2 fluence of about 1.5 megajansky-milliseconds and an energy near 10³⁴–10³⁵ erg at a distance of roughly 9–10 kiloparsecs. It sits just below the extragalactic population in energy.

The Theories

What Doesn't Add Up

• One engine or many? Repeaters and one-offs differ statistically in bandwidth and duration, and their host environments range from starbursts to dead galaxies. A single magnetar channel would need several formation pathways to cover this, and none has been observed producing an FRB except SGR 1935+2154.
• Why is the Galactic magnetar too faint? FRB 200428 was about ten times weaker than the least energetic extragalactic burst, and SGR 1935+2154 has not repeated at that level. The link is suggestive, not closed.
• What is the periodicity clock? A 16.35-day cycle with a frequency-dependent window has no agreed mechanism: orbital motion, precession and ultra-slow rotation have all been proposed.
• What are the persistent radio sources? FRB 121102 and a few other repeaters sit beside compact, steady radio emitters. Most localised FRBs have none.
• Do one-offs really not repeat? RBFLOAT's silence over six years is the best evidence yet, but a repeater with a very long cycle would look identical.
• What did JWST see at the RBFLOAT site? The faint near-infrared source is unexplained; if it is a red giant companion, it points toward binary models in an old population.

Latest Developments

The two years to September 2026 have been the most productive in the field's history. CHIME's three Outrigger telescopes came online in early 2025 and immediately delivered FRB 20250316A, localised to 13 parsecs in NGC 4141, the first FRB precise enough for JWST to resolve individual nearby stars; the radio and JWST papers appeared in The Astrophysical Journal Letters in August 2025 and were highlighted again in March 2026 when the no-repeat result was emphasised. The second CHIME/FRB catalogue, with 4,539 bursts and 83 repeaters, is now the reference population. In January 2025, FRB 20240209A became the first repeater in a quiescent elliptical host, at a record 40-kiloparsec offset. In January 2026, a FAST-led team published in Science the rotation-measure flare in FRB 220529A, the clearest evidence yet for a binary companion. DSA-110 has localised about 30 sources, ASKAP's coherent CRACO upgrade is finding bursts at higher sensitivity, and a 2026 forecast paper for the planned DSA/Chronoscope survey anticipates arcsecond localisations in the thousands, enough to use FRBs as a serious cosmological probe.

Sources

1. The discovery and significance of fast radio bursts — arXiv review (Lorimer, McLaughlin, Bailes), 2024
2. A Population of Fast Radio Bursts at Cosmological Distances — Science (Thornton et al.), 2013
3. Identifying the source of perytons at the Parkes radio telescope — arXiv / MNRAS (Petroff et al.), 2015
4. FRB 121102 Is Coincident with a Star-forming Region in Its Host Galaxy — Astrophysical Journal Letters, 2017
5. The First CHIME/FRB Fast Radio Burst Catalog — arXiv / Astrophysical Journal Supplement (CHIME/FRB Collaboration), 2021
6. CHIME delivers a second haul — Nature Astronomy research highlight, 2026
7. A fast radio burst associated with a Galactic magnetar — Nature (Bochenek et al., STARE2), 2020
8. Cosmic bursts unveil universe's missing matter — ScienceDaily / ICRAR on Macquart et al., 2020
9. A repeating fast radio burst source in a globular cluster — Nature (Kirsten et al.), 2022
10. Astronomers detect most distant fast radio burst to date — ESO on Ryder et al., 2023
11. First fast radio burst traced to old, dead, elliptical galaxy — Northwestern University, 2025
12. FRB 20250316A: A Brilliant and Nearby One-Off Fast Radio Burst Localized to 13 parsec Precision — arXiv / Astrophysical Journal Letters (CHIME/FRB Collaboration), 2025
13. Astronomers just found the source of the brightest fast radio burst ever — ScienceDaily / University of Toronto, 2026
14. A sudden signal flare reveals the hidden partner behind fast radio bursts — ScienceDaily / University of Hong Kong on Li et al., Science, 2026
15. The DSA/Chronoscope fast radio burst survey: forecasts and science overview — arXiv, 2026

The Marfa Lights

Physics students proved the famous lights are headlights on US-67. A few sightings a year still refuse to fit.

The Hook

In May 2004, twelve physics students from the University of Texas at Dallas drove to West Texas with lasers, walkie-talkies, a traffic counter, high-speed video and a chase car. They set up at the Marfa Lights viewing area on US-90 and watched the famous lights appear on the south-western horizon. Then they did the one thing the thousands of tourists before them had not: they parked a car twenty miles away on US-67, the highway to Presidio, and flashed its headlights.

At the viewing park, a Marfa light appeared.

Over four nights they counted every light and every vehicle. The two matched. The lights moved in a straight line that was, in fact, the road. A car overtaking another appeared as one orb passing through a second. Their report concluded that every light seen south-west of the park could be reliably attributed to automobile headlights.

The town has a bronze plaque at that park saying a cowboy named Robert Ellison first saw the lights in 1883, decades before cars. His own memoir, written in 1937, does not mention them. What the students explained, and what they did not, is the substance of this file.

The Facts

Marfa sits at about 4,688 feet in the Trans-Pecos, on a high plain called the Mitchell Flat that stretches south and west toward the Chinati Mountains and, beyond them, Mexico. The designated viewing site is a widened shoulder of US-90 about nine miles east of town, on the site of the former Marfa Army Airfield, where tens of thousands of personnel trained between 1942 and 1947. Observers there look south-west toward Chinati Peak. US-67 from Marfa to Presidio runs through that field of view, climbing and dipping through hills roughly fifteen to twenty miles away.

The lights as typically reported are distant points of white, yellow-orange, or occasionally red, green or blue light that brighten, dim, drift, split, merge and vanish, usually low above the desert but below the mesas behind them. Judith Brueske's 1989 guide distinguished them from ranch lights and headlights "primarily by their aberrant movements".

The documentary record is thin before the mid-twentieth century. Texas Monthly's Michael Hall, researching in 2006, found the earliest known press mention to be a February 1945 San Angelo Times item headlined "Ghost Light Appears in Marfa Area". The first magazine treatment was Paul Moran's "The Mystery of the Texas Ghost Light" in Coronet, July 1957. Elton Miles collected regional tales in 1976; a Wall Street Journal front page in 1984 spread the story nationally. The town built its first roadside viewing area and held its first Marfa Lights Festival in 1986. In 2003, using $720,000 in federal and state transportation money, it opened the current Marfa Mystery Lights Viewing Center, with mounted binoculars and plaques.

Two structured scientific efforts stand out. The 2004 UT Dallas Society of Physics Students study, cited above, attributed all lights observed south-west of the park over four nights to US-67 traffic. In May 2008 a Texas State University team led by electrical engineer Karl Stephan, working with the independent investigator James Bunnell, spent twenty nights at the viewing station with a Schmidt-Cassegrain telescope and a CCD spectrometer. They logged lights that "could have been mistaken for lights of unknown origin"; each was explained as headlights or small fires, and they showed that oxygen-absorption features in a light's spectrum could fix its distance to within six per cent at ranges beyond four kilometres. They concluded that genuinely anomalous lights are rare and that more monitoring was needed.

Bunnell, a retired aerospace engineer, is the source of most of that monitoring. After a sighting in 2000 he installed automated camera stations on ranches around the Flat, beginning in 2003, deliberately aimed away from US-67 and the Chinatis. His books catalogue 34 sightings from 1945 to 2008 and estimate an average of 9.5 anomalous lights on 5.25 nights per year, with the caveat that the cameras probably catch only half of what occurs.

Timeline

• 1883 – Robert Reed Ellison, a sixteen-year-old cowhand driving cattle through Paisano Pass, reportedly sees a flickering light he takes for an Apache campfire; the account survives only through family retelling.
• 1885 – Joe and Anne Humphreys are said to have seen the lights, per Robert and Judy Wagers' 2012 compilation.
• 1937 – Ellison writes a memoir of his life; his granddaughter and a journalist later find no mention of any lights in it.
• 1942–1947 – Marfa Army Airfield operates on the site of the future viewing area, training thousands of pilots.
• February 1945 – The San Angelo Times runs "Ghost Light Appears in Marfa Area", the earliest known press report.
• July 1957 – Coronet magazine publishes Paul Moran's "The Mystery of the Texas Ghost Light", the first widely cited written account.
• 1973 – Geologists Pat Kenney and Elwood Wright report two melon-sized lights approaching them on Nopal Road in the Mitchell Flat.
• 1976 – Elton Miles's Tales of the Big Bend collects the nineteenth-century stories and a rancher's photograph.
• 1984 – The Wall Street Journal puts the lights on its front page.
• 1986 – Marfa builds its first roadside viewing area and holds the first Marfa Lights Festival.
• 1989 – Judith Brueske publishes The Marfa Lights; a Japanese television team camps on the Flat for a week and sees nothing.
• 1992 – Herbert Lindee's "Ghost Lights of Texas" in Skeptical Inquirer argues for mundane explanations.
• 2000 – James Bunnell has the sighting that begins his long-term investigation.
• 2003 – The $720,000 Marfa Mystery Lights Viewing Center opens; Bunnell's first automated monitoring stations go live.
• May 2004 – UT Dallas Society of Physics Students conduct their four-night traffic-correlation study.
• 3 June 2005 – Two of Bunnell's stations record an extremely bright object glowing for over three hours during thunderstorms.
• May 2008 – Karl Stephan's Texas State team runs twenty nights of spectroscopy at the viewing station.
• 2009 – Stephan, Ghimire, Stapleton and Bunnell publish the spectroscopy study in the American Journal of Physics; Bunnell publishes Hunting Marfa Lights.
• 2011 – Stephan, Bunnell, Klier and Komala-Noor publish the 2005 event analysis in the Journal of Atmospheric and Solar-Terrestrial Physics.
• 2018 – BBC Travel profiles the "Ghost Lights" of Marfa; the viewing area remains a nightly draw.
• 29–31 August 2025 – The 38th annual Marfa Lights Festival is held at the Presidio County Courthouse.
• 4–6 September 2026 – The 39th festival is scheduled.

The Evidence

The evidence divides into three classes, and the classes do not agree.

The first is the routine viewing-park phenomenon, and it is effectively solved. The 2004 UT Dallas team used a traffic-volume monitor on US-67 and found that the frequency of lights south-west of the park tracked vehicle counts. The lights' apparent motion lay along a straight line matching the highway. Flashing a parked car's headlights produced a "Marfa light" on cue, and a passing car produced the classic merge-and-split. Hall replicated the observation with binoculars in 2006: each light switched on above and left of a radio-tower beacon, crept right, dimmed and vanished as the road dipped, and reappeared, tracing a descending line that ended near Marfa itself. Air over the Flat, which routinely swings 40 to 50 degrees Fahrenheit between day and night, bends and shimmers those distant headlights, which is why they appear to float and pulse.

The second class is the 2008 spectroscopic survey. Over twenty nights, Stephan's team combined a computer-logged azimuth-and-altitude mount, video and continuous spectra. Every candidate resolved into a continuum source at a distance consistent with vehicles or fires. The method matters as much as the result: molecular-oxygen absorption bands in a light's spectrum deepen with path length through air, giving a range estimate accurate to about six per cent beyond four kilometres. This is the first tool that can, in principle, distinguish a light on the highway from a light in the brush a mile away.

The third class is Bunnell's automated record, and it is the residue. His stations, pointed away from the highway, captured lights that rose, crossed the horizon and changed brightness in ways headlights on a fixed road cannot. The best-analysed case is 3 June 2005: an object recorded by two stations during a series of intense thunderstorms, emitting light for more than three hours. Optical triangulation put it on the ground about 28 kilometres from the farther station; radiometric analysis gave a peak visible-light output on the order of 10 kilowatts; and lightning strokes logged by the National Lightning Detection Network matched its origin in time and place. The 2011 paper does not identify the object. A three-hour, 10-kilowatt ground-level source triggered by lightning is consistent with a fire, but the authors stopped short of saying so.

Set against this is the absence of evidence where evidence should be. Ellison's memoir is silent. No pre-1945 newspaper report has been found. A military airfield patrolled nightly for five years produced no recorded anomaly. Judith Brueske's own definition of the lights, "distinguishable ... primarily by their aberrant movements", is a definition by exclusion. And Bunnell's count of roughly a dozen anomalous lights a year means that on almost every night, what a visitor sees is US-67.

The Theories

What Doesn't Add Up

• The 1883 origin. The Ellison sighting is repeated on the town's plaque, in the Texas State Historical Association handbook and in nearly every article, yet it rests on a family oral tradition first written down decades later, and Ellison's own 1937 memoir does not mention it. If the story is a later accretion, the lights have no documented history before 1945.
• Bunnell's residue. Automated cameras aimed away from the highway recorded lights that rose and crossed the horizon. These have not been independently replicated, but they have not been explained either.
• The 2005 object. Three hours, roughly 10 kilowatts, on the ground, tied to lightning: the published analysis stops at measurement and never names a cause.
• Local testimony. Ranchers who readily concede that the viewing-park lights are headlights, including the Mitchell family whose name the Flat carries, describe separate, rare events of large lights moving fast and low. They may be misperception; they have never been instrumented.
• The evolution of the story. Early accounts describe a single, unremarkable "ghost light"; modern accounts describe multiple fast-moving orbs. That drift is consistent with tourism and expectation shaping perception, but it has not been formally studied.

Latest Developments

No new field study of the Marfa lights has been published since the Stephan-Bunnell papers of 2009 and 2011, and the last two years have brought no instrumented sighting. The scientific thread that continues is Stephan's: in 2025 he and Richard Sonnenfeld of New Mexico Tech published an evaluation of the best video candidate for ball lightning their crowd-sourced database had received, and reported that peer review dismantled it, a result Texas State University publicised in February 2026 as a caution about how easily cameras produce convincing orbs. Their database of more than a thousand eyewitness reports remains open for submissions. In Marfa itself, the phenomenon is now primarily a cultural fixture: the Chamber of Commerce's 38th Marfa Lights Festival ran over Labor Day weekend 2025 at the Presidio County Courthouse, the 39th is scheduled for 4 to 6 September 2026, and the free viewing area on US-90, nine miles east of town, still fills each evening with visitors watching the Presidio highway without knowing it.

Sources

1. An Experimental Analysis of the Marfa Lights — Society of Physics Students, University of Texas at Dallas (Stolyarov, Klenzing, Roddy, Heelis), 2004
2. Spectroscopy applied to observations of terrestrial light sources of uncertain origin — American Journal of Physics (Stephan, Ghimire, Stapleton, Bunnell), 2009
3. Quantitative intensity and location measurements of an intense long-duration luminous object near Marfa, Texas — Journal of Atmospheric and Solar-Terrestrial Physics (Stephan, Bunnell, Klier, Komala-Noor), 2011
4. What Marfa Lights Probably Aren't — American Physical Society Texas Section meeting abstract (Stephan), 2006
5. The Truth Is Out There — Texas Monthly (Michael Hall), 2006
6. Skeptoid #38: The Marfa Lights: A Real American Mystery — Skeptoid (Brian Dunning), 2007
7. Marfa Lights — Handbook of Texas, Texas State Historical Association (Julia Cauble Smith)
8. Marfa lights — Wikipedia, accessed 2026
9. The mysterious 'Ghost Lights' of Marfa, Texas — BBC Travel, 2018
10. Scientists Can't Fully Explain These Strange Floating Lights in Texas — Vice, 2017
11. What Are the Marfa Lights — Live Science
12. Hunting Marfa Lights — Lacey Publishing (James Bunnell), 2009
13. Evaluation of video evidence of possible ball lightning — Quarterly Journal of the Royal Meteorological Society (Stephan and Sonnenfeld), 2025
14. Fact, fiction or a photo faux pas? Texas State researcher chases the glow behind ball lightning — Texas State University Newsroom, 2026
15. The Marfa Lights Festival shines a light on Labor Day weekend — Big Bend Sentinel, 2025

The Antikythera Mechanism

A 2,100-year-old bronze gearbox predicted eclipses; nothing like it appears again for a thousand years.

The Hook

On 17 May 1902, the archaeologist Valerios Stais was looking at a slab of corroded bronze the size of a cookbook, one of hundreds of objects pulled from a shipwreck the previous year. Inside the crust he saw the unmistakable profile of a toothed wheel. Nothing in the accepted story of ancient technology allowed for that. A gear-wheel with precisely cut teeth belonged to the clockmakers of medieval Europe, not to a cargo ship that went down off a Greek island around 60 BCE.

It took another seventy years and gamma-ray photographs to establish that the lump contained not one gear but a train of them, and another thirty years after that for a custom eight-tonne X-ray scanner to read the inscriptions hidden between the plates. The device modelled the Moon's irregular motion, counted down to eclipses across an 18-year cycle, and told its user which Panhellenic games fell that year. It is the only object of its kind ever found. Every attempt to explain it opens a larger question than the one it answers.

The Facts

The wreck lies off the north-east coast of Antikythera, a small island between Crete and the Peloponnese. Around Easter 1900, a crew of sponge divers from the island of Symi, under Captain Dimitrios Kontos, sheltered there from bad weather. Diver Elias Stadiatis came up from the seabed reporting a heap of bodies; they were bronze and marble statues. The Greek state, with the navy in support, ran a salvage operation from November 1900 until September 1902. It was the first large underwater archaeological recovery in history, done with helmet diving gear at depths that killed one diver and paralysed two others.

The mechanism was not noticed among the statues, glassware and pottery until Stais spotted the gear in May 1902. Today it survives as 82 fragments in the National Archaeological Museum in Athens: seven major pieces labelled A to G and 75 minor ones. Fragment A, the largest, holds most of the gearing. Thirty gear-wheels survive; reconstructions agree that the complete machine needed more, and most current models call for 37 or more.

What the device did is now largely settled, thanks to the Antikythera Mechanism Research Project's 2005 imaging campaign and the papers that followed in Nature in 2006 and 2008. A hand-turned input drove a front dial showing the position of the Sun and Moon in the zodiac against a 365-day Egyptian calendar ring. A small rotating ball displayed the phase of the Moon. On the back, a five-turn spiral tracked the 235-month Metonic cycle, and a four-turn spiral of 223 months tracked the Saros eclipse cycle, with glyphs marking predicted lunar and solar eclipses and the hour at which they would occur. Subsidiary dials tracked the 76-year Callippic cycle, the 54-year Exeligmos cycle, and a four-year dial naming the games of the Olympiad circuit, including Olympia, Nemea, Isthmia, Pythia and the Naa at Dodona.

The mechanism's most sophisticated trick is a pin-and-slot coupling between two gears mounted on slightly offset axes. As one wheel drives the other, the pin slides in and out of the slot, so the output speeds up and slows down once per cycle. This reproduces the first lunar anomaly, the variation in the Moon's speed later described by Hipparchus, to within a fraction of a percent.

Inscriptions cover the surviving plates: a parapegma, or star calendar, on the front; user instructions on the back; and a description of a planetary display. Letter-forms place the engraving in the middle of the second century BCE, and the wreck's coins, minted at Pergamon and Ephesus between roughly 76 and 67 BCE, put the sinking around 70 to 60 BCE.

Timeline

• c. 205 BCE: Astronomical epoch on which the Saros eclipse dial appears to have been calibrated (Carman and Evans 2014; Freeth 2014). Disputed as a construction date.
• c. 150 to 100 BCE: Probable construction window based on letter-forms of the inscriptions.
• c. 70 to 60 BCE: The cargo ship sinks off Antikythera, dated by coins and ceramics.
• Easter 1900Symi sponge divers under Dimitrios Kontos discover the wreck; Elias Stadiatis surfaces with a bronze arm.
• November 1900 to September 1902: Greek navy-supported salvage recovers statues, pottery and glass; one diver dies.
• 17 May 1902Valerios Stais identifies a gear-wheel in a corroded bronze fragment at the National Archaeological Museum.
• 1953 and 1976: Jacques Cousteau surveys the wreck and then excavates it, recovering coins and small finds.
• 1959Derek de Solla Price publishes "An Ancient Greek Computer" in Scientific American.
• 1971 to 1974: Price and Charalambos Karakalos take gamma radiographs; Price publishes "Gears from the Greeks" (1974), reading a Metonic gear train.
• 1990s to 2005: Michael Wright and Allan Bromley apply linear tomography; Wright builds the first working planetarium reconstruction.
• 2005The Antikythera Mechanism Research Project brings an eight-tonne microfocus CT scanner and surface-imaging rigs to Athens.
• 30 November 2006Freeth et al. publish "Decoding the ancient Greek astronomical calculator" in Nature, doubling the readable text and establishing the eclipse dial and the pin-and-slot.
• 31 July 2008Freeth, Jones, Steele and Bitsakis publish the Olympiad dial and Corinthian month names in Nature.
• 2014Carman and Evans, and separately Freeth, independently date the eclipse dial's start to 205 BCE.
• 2016The full inscription corpus, roughly 3,400 characters, is published in a special issue of Almagest; a skeleton nicknamed "Pamphilos" is recovered from the wreck.
• 12 March 2021The UCL team led by Tony Freeth publishes a complete front-dial "Cosmos" model in Scientific Reports.
• June 2022Divers recover a marble head believed to belong to the Herakles statue raised in 1901.
• June to July 2024: Glasgow gravitational-wave researchers Graham Woan and Joseph Bayley publish a Bayesian hole-count of 354 or 355 for the calendar ring.
• April 2025Szigety and Arenas post a simulation arguing the device would have jammed within about four months if the gears were as imprecise as measured.
• 23 May to 20 June 2025: The Swiss School of Archaeology's five-year campaign concludes by raising a connected section of the ship's hull.

The Evidence

The physical evidence is small, dense and unforgiving. Fragment A measures roughly 18 by 15 centimetres. Its main drive wheel, about 13 centimetres across, is the largest gear in the assembly. The gears are cut from a single sheet of low-tin bronze around two millimetres thick, with triangular teeth. Tooth counts recovered from the CT data include 223, the number of synodic months in a Saros; 127, half of 254, the number of sidereal lunar months in 19 years; and 53, a prime that appears twice in the lunar train and makes sense only as a factor in a deliberate ratio. These numbers are the strongest single argument that the machine was designed from astronomical theory rather than assembled by trial.

The 2005 scans used an X-Tek "Bladerunner" microfocus CT unit shipped to Athens, with resolution fine enough to read letters engraved on plates buried inside the corrosion, while Hewlett-Packard's polynomial texture mapping captured surface detail under controlled lighting. The 2006 Nature paper reported that this doubled the legible text; the 2016 Almagest corpus brought the total to roughly 3,400 characters, from what was probably a much longer original.

The calendar ring on the front carries a ring of small holes that pinned it in place and let the user rotate it to correct the Egyptian year against the seasons. Only a fraction of the ring survives, in six pieces. In 2020 a group including Chris Budiselic measured the surviving holes and argued the full ring had about 354 holes, matching a lunar year rather than the 365-day Egyptian calendar. In 2024 Woan and Bayley at the University of Glasgow reanalysed the hole positions using Markov chain Monte Carlo and nested sampling, methods built for LIGO gravitational-wave data. Their result: the ring most likely had 354 or 355 holes, drilled on a circle of radius 77.1 millimetres with an uncertainty of about a third of a millimetre, and with an average radial scatter between holes of just 0.028 millimetres. That level of precision is a data point in itself about the workshop that made it.

Two other evidence lines are contested. First, the accuracy of the mechanism as a working instrument: Mike Edmunds, who chaired the research project, measured gear errors from the scans, and in 2025 Esteban Szigety and Gustavo Arenas of the Universidad Nacional de Mar del Plata built a simulation combining those errors with Alan Thorndike's analysis of triangular-tooth motion. They found the triangular teeth alone were harmless, but the measured manufacturing errors would cause jamming or disengagement before the Sun pointer had advanced about 120 days. The authors themselves think it more likely that corrosion has distorted the gears than that the original device was unusable.

Second, the shipwreck context. The 2012 to 2019 "Return to Antikythera" project, co-directed by Brendan Foley with the Greek Ephorate of Underwater Antiquities, recovered a human skeleton in 2016 and a bronze statue arm in 2017. The 2021 to 2025 Swiss School of Archaeology campaign under Lorenz Baumer and Angeliki Simosi raised, in 2025, three outer planks still joined to an internal frame, confirming shell-first construction; preliminary dating of the wood gives around 235 BCE, and the species are elm and oak. None of this touches the mechanism directly, but it fixes the ship as a large, expensive vessel carrying luxury cargo west, most likely toward Italy.

The Theories

What Doesn't Add Up

• Why is there only one? A device this refined implies predecessors and successors. Nothing comparable survives from the ancient world, and geared astronomical instruments do not reappear until Byzantine and Islamic devices centuries later. Either the tradition was tiny and left no trace, or the archaeological record has swallowed it.
• Who made it and where? Corinthian month names point west, Rhodian pottery and Hipparchan theory point east, and the cargo was headed toward Italy. No workshop, no maker's name, no sponsor is known.
• What was the front display? The inscriptions describe planets, but the gearing to show them is missing. The 2021 UCL model is elegant and consistent, and unverified by any surviving part.
• Did it actually work? The 2025 simulation argues the measured gear geometry would jam it within four months. The hole-count precision argues for a far better maker. The two results cannot both describe the original machine.
• Why does the Saros dial calibrate to 205 BCE? That epoch precedes the inscription letter-forms by half a century. Either the maker chose an old epoch deliberately or the construction date needs revisiting.
• Is the wreck a single ship? Some researchers have raised the possibility of a second vessel at the site. The 2025 hull section confirms one large ship, but the full extent of the debris field has never been mapped to the edges.

Latest Developments

The last three years have pulled in opposite directions. In 2024 the Glasgow analysis settled a long-running argument about the calendar ring, finding strong evidence for 354 or 355 holes and therefore a Greek lunar calendar rather than the Egyptian solar year, and it did so with methods borrowed from gravitational-wave astronomy. In April 2025 the Argentine simulation study reopened the question of whether the device could physically run, concluding that the measured tooth errors would stall it within roughly 120 days while cautioning that the measurements probably reflect corrosion rather than original workmanship. Meanwhile, the Swiss School of Archaeology's five-year excavation ended in June 2025 with the first recovery of an articulated section of the hull, wood dated to about 235 BCE, and new pottery finds spread across two zones of the site. Reconstruction work continues in 2026, including revised front-dial designs based on the lunar calendar reading; neither modern excavation campaign has recovered a new fragment of the mechanism itself.

Sources

1. Decoding the ancient Greek astronomical calculator known as the Antikythera Mechanism — Nature, 2006
2. Calendars with Olympiad display and eclipse prediction on the Antikythera Mechanism — Nature, 2008
3. A Model of the Cosmos in the ancient Greek Antikythera Mechanism — Scientific Reports, 2021
4. Gravitational wave researchers cast new light on Antikythera mechanism mystery — University of Glasgow, 2024
5. An improved calendar ring hole-count for the Antikythera mechanism — arXiv (Woan and Bayley), 2024
6. The Impact of Triangular-Toothed Gears on the Functionality of the Antikythera Mechanism — arXiv (Szigety and Arenas), 2025
7. Antikythera Mechanism's intricate gears: Simulations reveal potential jamming issues — Phys.org, 2025
8. On the epoch of the Antikythera mechanism and its eclipse predictor — Archive for History of Exact Sciences (Carman and Evans), 2014
9. The Epoch Dates of the Antikythera Mechanism — ISAW Papers (Jones), 2020
10. Gears from the Greeks: The Antikythera Mechanism — American Philosophical Society (Price), 1974
11. Ancient Ship Remains Rescued from Antikythera Wreck — Swiss School of Archaeology in Greece, 2025
12. Antikythera shipwreck reveals ancient shipbuilding secrets in 2025 excavation — Archaeology News, 2025
13. Human Skeleton Found on Famed Antikythera Shipwreck — Scientific American, 2016
14. Bronze Arm Found at Antikythera Shipwreck — Live Science, 2017
15. On this day in 1902, the Antikythera Mechanism was discovered — Greek City Times, 2021

The Great Attractor

Our galaxy falls at 600 km/s toward a point hidden behind the Milky Way; the mass to explain it is missing.

The Hook

Subtract the rotation of the Earth, the orbit of the Sun around the Milky Way, and the Milky Way's own dance with Andromeda, and one motion remains. The entire Local Group of galaxies is moving at about 630 kilometres per second relative to the afterglow of the Big Bang. That number is not in dispute. It is stamped on the cosmic microwave background as a dipole, a hot patch ahead of us and a cold one behind, first measured in 1977.

The trouble is the direction. The motion points into the southern sky, low along the plane of our own galaxy, where dust and stars block roughly a fifth of the extragalactic universe from view. When a team of seven astronomers went looking for the cause in the mid-1980s, they found galaxies for hundreds of millions of light-years around us all streaming the same way, toward a region they could not see. They called it the Great Attractor. Forty years on, the flow is measured to better precision than ever. What sits at the end of it is still argued in the literature, and in 2026 two peer-reviewed groups reached opposite conclusions.

The Facts

The story begins with the cosmic microwave background. In 1977 George Smoot and colleagues flew a differential radiometer on a U-2 aircraft and measured a dipole in the CMB temperature. The simplest reading is Doppler: our Sun moves at about 370 km/s relative to the radiation, and after correcting for the Sun's orbit in the Milky Way and the Milky Way's motion in the Local Group, the Local Group itself moves at 631 plus or minus 20 km/s toward Galactic longitude about 276 degrees and latitude about 30 degrees.

Through the early 1980s, astronomers looked for the local overdensity that could produce such a velocity. In 1986 Per Lilje and colleagues found a tidal distortion of the velocity field in the Local Supercluster whose dominant axis pointed toward the Hydra-Centaurus clusters. The decisive measurement came from a collaboration nicknamed the Seven Samurai: David Burstein, Roger Davies, Alan Dressler, Sandra Faber, Donald Lynden-Bell, Roberto Terlevich and Gary Wegner. They measured distances to about 400 elliptical galaxies using a new relation between a galaxy's size, brightness and internal velocity dispersion, then compared those distances with redshifts to obtain each galaxy's peculiar velocity, its departure from pure Hubble expansion.

Dressler and colleagues reported in 1987 that galaxies out to about 6,000 km/s share a coherent bulk flow of several hundred kilometres per second, and that the Hydra-Centaurus region itself is moving, so the source must lie beyond it. Lynden-Bell and colleagues in 1988 fitted an attractor model instead, with peculiar velocity falling off with distance from a centre at Galactic coordinates roughly l 307, b 9, at 43.5 h-inverse megaparsecs, a recession velocity near 4,350 km/s. Dressler coined the name. The implied mass was of order 5 times 10 to the 16 solar masses, spread across a region a few hundred million light-years wide.

The centre sat close to the Zone of Avoidance, the band about 20 degrees wide along the Galactic plane where extinction hides distant galaxies. In 1996 Renée Kraan-Korteweg and collaborators, surveying that band, identified the Norma cluster, ACO 3627, at l 325, b minus 7, receding at 4,871 km/s, roughly 68 megaparsecs or 220 million light-years away. It is the richest cluster in the region, comparable in mass to Coma, and is now widely treated as the core of the Great Attractor, though it lies about 18 degrees from the Lynden-Bell centre.

Then the picture widened. Behind the Great Attractor, at about 650 million light-years in Centaurus, lies the Shapley Supercluster, first noticed by Harlow Shapley in 1930 and named by Somak Raychaudhury in 1989: more than 8,000 galaxies, the most massive concentration known within a billion light-years. In 2014 Brent Tully, Hélène Courtois, Yehuda Hoffman and Daniel Pomarède mapped the watershed of galaxy flows from 8,000 peculiar velocities and defined Laniakea, a supercluster 500 million light-years across containing about 100,000 galaxies, with the Great Attractor as its flat-bottomed gravitational valley.

Timeline

• 1930Harlow Shapley notes an unusually dense cloud of galaxies in Centaurus, later the Shapley Supercluster.
• 1977Smoot, Gorenstein and Muller measure the CMB dipole from a U-2 aircraft; the Local Group is moving.
• 1982Aaronson and colleagues detect departures from smooth Hubble flow but cannot locate the source.
• 1986Lilje and colleagues find a tidal field in the Local Supercluster pointing at Hydra-Centaurus.
• 1987Dressler and the Seven Samurai report a coherent large-scale streaming motion from 400 elliptical galaxies.
• 1988Lynden-Bell and colleagues fit an attractor model; the Great Attractor gets its coordinates and its name.
• 1989Raychaudhury names the Shapley Supercluster from the APM plate survey.
• Early 1990s: Surveys of spiral galaxies beyond the attractor find no clear backside infall, opening a long dispute over its mass and extent.
• 1996Kraan-Korteweg and colleagues identify the Norma cluster (ACO 3627) as the richest cluster hidden in the Great Attractor region.
• 2004Woudt, Kraan-Korteweg and others determine the Norma cluster's distance and peculiar velocity using the Fundamental Plane.
• 4 September 2014Tully, Courtois, Hoffman and Pomarède publish Laniakea in Nature; the Great Attractor becomes its basin floor.
• 2017Hoffman, Pomarède, Tully and Courtois publish the Dipole Repeller in Nature Astronomy; a void pushes as hard as the attractor pulls.
• 2017Kraan-Korteweg and colleagues report the Vela Supercluster at about 18,000 km/s, straddling the Galactic plane.
• 2023Cosmicflows-4 (Tully and colleagues) compiles distances for 55,877 galaxies; Watkins and colleagues find a bulk flow of 419 km/s at 200 h-inverse megaparsecs, in tension with the standard model.
• 2024Valade, Hoffman, Libeskind and colleagues map basins of attraction from Cosmicflows-4; Laniakea may be part of the Shapley basin.
• 2024A blind MeerKAT HI survey begins to map the Vela region through the densest part of the Zone of Avoidance.
• January 2026Stiskalek, Desmond, Hudson and colleagues argue from constrained simulations that the classical Great Attractor is a velocity-field artefact.
• March 2026Hollinger, Courtois, Kraan-Korteweg and colleagues report Vela as a mass concentration rivalling Shapley.
• April to May 2026: Dressler and Monson report surface-brightness-fluctuation distances showing a flow converging at about 70 megaparsecs, consistent with the original Great Attractor.

The Evidence

The core evidence is a velocity field, not an image. Astronomers measure a galaxy's redshift, which combines cosmic expansion with the galaxy's own motion, and separately estimate its distance by a redshift-independent method. The difference is the peculiar velocity along the line of sight. Distance estimators are the weak link: the Tully-Fisher and Fundamental Plane relations carry errors of 20 to 25 percent per galaxy, so flows are recovered by averaging thousands of objects. Cosmicflows-4, published in 2023, is the largest such compilation, with 55,877 galaxies in 38,065 groups and eight distance methods. It implies a Hubble constant of 74.6 km/s per megaparsec with a statistical error of 0.8 and a systematic uncertainty near 3.

Applied to this catalogue, the minimum-variance method of Watkins and colleagues gives a bulk flow of 419 plus or minus 36 km/s within a sphere of radius 200 h-inverse megaparsecs, pointing to l 298, b minus 8. The standard cosmological model predicts a much smaller flow on that scale; the authors put the tension at 4.8 sigma. If the whole local volume is drifting, no single nearby mass can be responsible.

The Norma cluster provides the one hard anchor inside the Zone of Avoidance. Its mean velocity of 4,871 km/s and dispersion of 925 km/s from 296 members, measured by Woudt and colleagues in 2008, put its mass in the range of the Coma cluster. Near-infrared Fundamental Plane distances give about 68 megaparsecs and a peculiar velocity consistent with zero, which is what one expects of a cluster sitting at or near the bottom of a potential well.

The 2017 Dipole Repeller analysis reconstructed the three-dimensional flow from Cosmicflows-2 and found the Local Group's motion is best described as a push from an underdense region roughly 15,000 km/s away, opposite Shapley on the sky, combined with a pull from the Shapley concentration, the two contributing about equally. The Great Attractor, in this reconstruction, is a way-station rather than the destination.

Two 2026 datasets sharpen the contradiction. Dressler and Andrew Monson used the FourStar near-infrared camera on the 6.5-metre Magellan-Baade telescope to obtain surface-brightness-fluctuation distances, accurate to about 5 percent, for 66 galaxies with velocities between 2,000 and 5,000 km/s. They report a strong flow across roughly a steradian of sky, peaking near 1,000 km/s and converging to zero at about 70 megaparsecs from the Local Group. That matches the original 1988 attractor model in scale and lines up with the CMB dipole. They argue that the much larger Cosmicflows-4 bulk flows are artefacts of less precise distance estimators.

Separately, Hollinger, Courtois, Kraan-Korteweg, Mould and Rajohnson combined 65,518 peculiar-velocity distances from an extended CF4++ catalogue with 8,283 new redshifts within 10 degrees of the southern Galactic plane, including 2,176 HI redshifts from MeerKAT covering the innermost 3-degree strip of the Zone of Avoidance for the first time. Their reconstruction gives the Vela Supercluster, at about 189 h-inverse megaparsecs, a total mass of 33.8 times 10 to the 16 solar masses and a radius of 70 h-inverse megaparsecs, rivalling Shapley and exceeding the mass they associate with Laniakea and the Great Attractor region.

The Theories

What Doesn't Add Up

• The mass budget does not close. The identified clusters in the Great Attractor region, Norma included, fall well short of the 5 times 10 to the 16 solar masses the 1988 model needed to produce the observed infall.
• The centre keeps moving. Lynden-Bell put it at l 307, b 9; Norma is at l 325, b minus 7; the 2026 Manticore analysis finds convergence near Abell 3565, which excludes Norma; Dressler and Monson find a convergence radius of 70 megaparsecs. These are not all the same object.
• Two 2026 results contradict each other. One team, using 5 percent distances to 66 galaxies, confirms a compact, steradian-sized flow matching the original Great Attractor. Another, using constrained simulations against catalogues of tens of thousands of galaxies, concludes no such dominant structure exists.
• The bulk-flow scale is unresolved. Cosmicflows-4 analyses report flows of over 400 km/s extending 200 h-inverse megaparsecs at 4.8 sigma tension with the standard model, while precision distance work sees the flow die out at a third of that distance.
• Roughly 20 percent of the sky remains unmapped for peculiar velocities. Every reconstruction interpolates across the Zone of Avoidance, and every new survey of it, from Norma to Vela, has found mass no model had placed there.
• Even summed, known structures within 155 h-inverse megaparsecs account for only about 72 percent of the Local Group's velocity in the 2026 Manticore simulations, with a 38-degree offset in direction.

Latest Developments

The last three years have turned a slow debate into an open contest. Cosmicflows-4 (2023) delivered distances for nearly 56,000 galaxies and, in the hands of Watkins and colleagues, a bulk flow far larger than Lambda-CDM predicts. In 2024 Valade and colleagues used the same catalogue to redraw the basins of attraction, finding a slight preference for Laniakea to sit inside the much larger Shapley basin. In January 2026 Stiskalek and colleagues published constrained simulations concluding that no single structure, the Great Attractor included, dominates the Local Group's motion. In March 2026 the first hybrid reconstruction combining CF4++ velocities with MeerKAT HI redshifts inside the Zone of Avoidance promoted the Vela Supercluster to a mass rivalling Shapley. Then in April 2026 Alan Dressler, one of the original Seven Samurai, returned with Andrew Monson and 5-percent surface-brightness-fluctuation distances, reporting a steradian-sized flow converging at about 70 megaparsecs and aligned with the CMB dipole, much as proposed in 1988. The case is not closing. It is reopening.

Sources

1. Spectroscopy and photometry of elliptical galaxies: A large-scale streaming motion in the local universe — Astrophysical Journal Letters (Dressler et al.), 1987
2. Donald Lynden-Bell: A Biographical Memoir — arXiv / National Academy of Sciences, 2020
3. The Norma cluster (ACO 3627) I: A dynamical analysis of the most massive cluster in the Great Attractor — Monthly Notices of the RAS (Woudt et al.), 2008
4. Norma cluster (ACO 3627) III: The distance and peculiar velocity via the near-infrared Ks-band Fundamental Plane — Monthly Notices of the RAS, 2014
5. The Laniakea supercluster of galaxies — Nature (Tully, Courtois, Hoffman, Pomarède), 2014
6. The dipole repeller — Nature Astronomy (Hoffman et al.), 2017
7. Discovery of a supercluster in the Zone of Avoidance in Vela — MNRAS Letters (Kraan-Korteweg et al.), 2017
8. Cosmicflows-4 — Astrophysical Journal (Tully et al.), 2023
9. Analysing the large-scale bulk flow using Cosmicflows-4: increasing tension with the standard cosmological model — Monthly Notices of the RAS (Watkins et al.), 2023
10. Identification of basins of attraction in the local Universe — Nature Astronomy (Valade et al.), 2024
11. Revealing hidden structures in the Zone of Avoidance: a blind MeerKAT HI survey of the Vela Supercluster — Monthly Notices of the RAS, 2024
12. Revisiting the Great Attractor: The Local Group's streamline trajectory, cosmic velocity and dynamical fate — arXiv (Stiskalek et al.), 2026
13. Hidden Vela Supercluster Revealed by First Hybrid Redshift and Peculiar Velocity Reconstruction — arXiv (Hollinger et al.), 2026
14. Return to the Great Attractor: Strong Evidence for a Steradian-sized Flow Converging at 70 Mpc — arXiv (Dressler and Monson), 2026
15. Shapley Supercluster — ESA Science and Technology, undated

The Carroll A. Deering

A five-masted schooner ran onto Diamond Shoals with every sail set, food on the stove and no one aboard

The Hook

At first light on Monday, 31 January 1921, a surfman on lookout at Coast Guard Station 183, Cape Hatteras, saw a large five-masted schooner sitting motionless on the outer Diamond Shoals with all her canvas set. No distress flags flew. No rockets had been seen during the night. When a surfboat fought its way to within a quarter of a mile of her, the crew could make out two empty davits with their falls trailing over the side, a ladder hanging down the hull, and no movement on deck.

It took four days of storm before anyone could board. Coffee, soup and spareribs stood in the galley. The captain's cabin was in disarray. The ship's papers, chronometer, instruments, log and two boats were gone. So were eleven men. None of them, nor either boat, has ever been found.

The Facts

The Carroll A. Deering was launched on 4 April 1919 at the G. G. Deering Company yard in Bath, Maine, the last of roughly a hundred vessels the company built and among the last large wooden cargo schooners built anywhere. She was 255 feet long and 44 to 45 feet in beam. Her tonnage is reported inconsistently: 1,879 tons in most modern sources, 2,114 tons in Karl Baarslag's 1937 Coast Guard account.

Her final voyage began in late August 1920 when she left Virginia with coal for Rio de Janeiro under Captain William H. Merritt, with Merritt's son Sewall as mate. Merritt fell ill within days and both Merritts left her at Lewes, Delaware. The company brought Captain Willis B. Wormell, 66 and effectively retired, out to take command; Charles B. McLellan was signed as first mate. She sailed again on 8 September 1920, delivered her coal at Rio and departed on 2 December in ballast for Hampton Roads via Barbados.

The crew, per the federal investigators cited by Baarslag, consisted of an American mate and engineer, a Finnish boatswain, a Black steward and six Danish sailors. With Wormell that makes eleven; sources variously give ten, eleven or twelve.

What happened on the way home rests on statements gathered later. In Rio, Wormell reportedly told a friend, Captain Goodwin, that he distrusted his crew except for the engineer, Herbert Bates. In Barbados, McLellan got drunk, complained to Captain Hugh Norton of the schooner Snow that Wormell undercut his discipline and that he had to do the navigating because of the captain's failing eyesight, and was overheard saying he would "get" the old man before Norfolk. He was arrested; Wormell bailed him out. Baarslag adds that Wormell told the Barbados agents he was in poor health. The Deering left Barbados on 9 January 1921.

On 28 or 29 January (sources split on the date) she passed the Cape Lookout Lightship. A tall, thin, red-haired man with a foreign accent hailed it through a megaphone, said the ship had lost both anchors in a storm off Cape Fear, and asked that the owners be notified. Keeper Jacobson could not relay the message because his radio was out. He later said the crew were "milling around" on the quarterdeck, where seamen were not normally allowed, and that the man who hailed him did not look like an officer. A lightship engineer photographed her as she passed; it is the last known image of her under way. At about 5:45 p.m. on 30 January the steamer Lake Elon saw her steering, in the National Park Service's words, "a peculiar course" toward the shoals. No one hailed her.

On 4 February 1921 the wrecking tug Rescue, working with the cutter Manning, reached the wreck. Captain James Carlson of the Rescue was first aboard. The wheel was smashed, the binnacle stove in, the rudder disengaged from its stock. Both running lights were still burning, as were two red lights in the rigging. Three half-starved cats were found. Charts lay in the captain's bathroom; entries in Wormell's hand stopped on 23 January and continued in another hand. Beds were unmade, dishes unwashed, three pairs of boots left behind. Stores, signal flags and gear were untouched. Baarslag records "no evidence of foul play," while conceding that four days of sea and sand would have erased any.

Timeline

• 4 April 1919Carroll A. Deering launched at Bath, Maine.
• 8 September 1920After Captain Merritt falls ill, sails for Rio under Captain Willis B. Wormell.
• 2 December 1920Departs Rio in ballast for Hampton Roads via Barbados.
• 9 January 1921Leaves Barbados.
• 23 January 1921Last chart entry in Wormell's hand.
• 25 January 1921Last radio call from the SS Hewitt; never heard from again.
• 28 or 29 January 1921: Deering hails the Cape Lookout Lightship, reports both anchors lost.
• 30 January 1921Lake Elon sights the Deering steering toward Diamond Shoals.
• 31 January 1921Surfman C. P. Brady sights the schooner aground.
• 4 February 1921Rescue and Manning reach the wreck; Carlson boards and confirms abandonment.
• 21 March 1921Seminole destroys the drifting stern section.
• April 1921Christopher Columbus Gray reports a message in a bottle claiming the Deering was captured.
• May-June 1921: Wormell's family petitions Washington; a five-department investigation begins under Lawrence Richey.
• 21-22 June 1921: Newspapers nationwide report piracy suspected.
• 20 August 1921Richey confronts Gray with handwriting reports; Gray admits forging the note.
• 1922Investigation ends without an official finding.
• September 1955Hurricane Ione moves the bow section from Ocracoke to Hatteras.
• 30-31 January 2021: Centennial marked by Cape Hatteras National Seashore and the Graveyard of the Atlantic Museum.

The Evidence

The state of the ship. Her keel was buried deep in the shoal - Bland Simpson, whose 2002 book is the most detailed reconstruction, puts it at fourteen feet - which the surfmen took to mean she had struck under full sail with great force. Coast Guard Captain R. L. Gaskill of Station 184 argued that fourteen set sails proved the ship was manned when she struck. Captain Ballance of the Cape Hatteras station told the federal agent the opposite: that the crew "abandoned her after taking everything of value" and ran her onto the shoals intentionally.

What was missing and what was not. The two boats, the anchors, all papers, the chronometer, instruments, log and the crew's good clothing were gone. Stores, flags, the medicine chest, furniture and sails were intact and sold at auction. Investigators read the pattern as an orderly departure rather than a panic. Simpson records a nine-pound sledgehammer lying near the smashed wheel; whether the damage was done by people or by four days of surf has never been settled.

The red lights. Two red lights in the rigging, still burning on 4 February, signal a vessel not under command. They fit a crew that knew the anchors were gone and the steering was failing; they also fit someone wanting the ship to look disabled.

The lightship testimony. Jacobson's account is the only eyewitness description of the crew after Barbados. Lightship crew later shown photographs of Wormell said the man at the rail was not him. Jacobson also reported an unidentified steamer that passed shortly after, ignored his hail and appeared to have canvas over her name.

The message in a bottle. Gray's note read: "Deering captured by oil burning boat something like chaser. Taking off everything Handcuffing crew. Crew hiding all over ship. No chance to make escape. Finder please notify headqts. Deering." Wormell's family identified the hand as Bates's. Handwriting experts disagreed sharply. In August 1921 Treasury and Navy examiners, working from Gray's own Navy service file, concluded he wrote it. Confronted by Richey on 20 August, Gray confessed; his motive, in Simpson's account, was to embarrass the Cape Hatteras light station staff and improve his own chances of a job there. He was never prosecuted and died in Buxton in 1952.

The Richey files. The Lawrence Richey Papers at the Herbert Hoover Presidential Library preserve the investigation's correspondence, reports and photographs. The library's summary states that Richey and Hoover concluded mutiny was most likely but left the case open.

The Theories

What Doesn't Add Up

The record is contradictory even on basics. The Cape Lookout sighting is 28 January in the Coast Guard photograph caption and 29 January in the Park Service account. The crew is ten, eleven or twelve. The tonnage is 1,879 or 2,114. The petitioner who moved Hoover is Wormell's wife Lula in the Park Service account and his daughter in the Hoover Library's. Small discrepancies, but they show how much was written down by people who were not there.

The larger problems are structural. A mutinous crew had no reason to abandon a sound ship. A crew fleeing a storm should have left some trace on a coast searched for weeks. Pirates had no reason to leave everything saleable and take only the paperwork. The removal of the log and instruments is the one deliberate act everyone agrees on, and it is exactly what a mutinous crew, a fleeing crew and a boarding party would each have wanted to do.

Two details resist every theory. The chart entries changed hands on 23 January, a week before the grounding, which suggests Wormell was no longer navigating - dead, incapacitated or deposed - long before North Carolina. And the ship carried the "not under command" lights while under full sail, an odd combination for anyone who intended either to save her or to sail her away.

Finally, the investigation destroyed its own evidence. The Deering was stripped on 4 February, sold off and blown up within weeks.

Latest Developments

The centennial passed quietly. On 30 January 2021 Cape Hatteras National Seashore and the Graveyard of the Atlantic Museum posted commemorations online; the museum's director, Joe Schwarzer, had planned a virtual event inviting the public to weigh the evidence. The museum in Hatteras Village holds the ship's bell and capstan, the captain's desk, a flask and a lantern, and has built Deering timbers into its gift shop facade. Its administrator told McClatchy in 2021 that a Deering clock in a local family's possession, silent for decades, began chiming on 22 August 2020, the anniversary of the final departure for Rio.

Scholarship has not moved much. Bland Simpson's Ghost Ship of Diamond Shoals (UNC Press, 2002; paperback 2005), built on FBI reports, ships' logs and correspondence, remains the standard account. James D. Charlet's Shipwrecks of the Outer Banks (2020) added a day-by-day reconstruction of the rescue attempts from station and cutter logs. The Library of Congress published a survey of 1921 press coverage in January 2023, and the National Coast Guard Museum in New London republished Baarslag's 1937 chapter in October 2024 ahead of an exhibit on aids to navigation. Regional press revisited the case in 2024 and 2025. None of this produced new evidence, and no comprehensive re-examination of the Richey Papers has been published.

The wreck itself is gone. The bow section that lay off Ocracoke for 34 years was carried to Hatteras by Hurricane Ione in September 1955 and is now buried; the Park Service states that no remains are visible on the beach. The case is open because there is nothing left to close it with.

Sources

1. The Ghost Ship of the Outer Banks — National Park Service, Cape Hatteras National Seashore, 2015
2. The Wreck of the Carroll A. Deering — Herbert Hoover Presidential Library (National Archives), 2016
3. The Mysterious Disappearance of Ghost Ship Carroll A. Deering's Crew — Library of Congress, 2023
4. The Coast Guard responds to ghostship Carroll A. Deering — National Coast Guard Museum (adapted from Karl Baarslag, 1937), 2024
5. The Carroll A. Deering, "Ghost Ship" — North Carolina Department of Natural and Cultural Resources, 2016
6. The Ghost Ship Caroll A. Deering: Still an Outer Banks Mystery — NCpedia / State Library of North Carolina, 2014
7. Ghost Ship of Diamond Shoals: The Mystery of the Carroll A. Deering — University of North Carolina Press, 2005
8. The Attempted Rescues of the Famous Outer Banks "Ghost Ship," the Carroll A. Deering — Island Free Press, 2019
9. January 31 Marks the 100th Anniversary of the Infamous "Ghost Ship" of the OBX — Island Free Press, 2021
10. 'Ghost ship' with a cat - but no crew - wrecked on Outer Banks 100 years ago this week — WBTV / Charlotte Observer, 2021
11. The mystery of the 'ghost ship' Deering — The Charlotte Ledger, 2025
12. Ghost Ship of the Outer Banks: What really happened to the Carroll A. Deering? — Cardinal & Pine, 2024
13. The Curious Case of The Carroll A. Deering — Fishermen's Voice, 2020
14. Carroll A. Deering — Wikipedia, 2026
15. SS Hewitt — Wikipedia, 2026

The Pioneer Anomaly

Two 1970s probes drifted thousands of kilometres off course; a rescued tape archive finally explained why

The Hook

In August 1998, six researchers led by John D. Anderson of NASA's Jet Propulsion Laboratory published a short paper in Physical Review Letters with a cautious title and an uncomfortable conclusion. Radio tracking of Pioneer 10 and Pioneer 11, the first probes ever sent past Jupiter, showed both vehicles experiencing a tiny, constant, unexplained acceleration back toward the Sun. The magnitude was roughly 8.5 × 10⁻⁸ cm/s², about ten billion times weaker than gravity at Earth's surface. Two independent software codes had found it. A list of conventional causes had been ruled out.

The effect was small enough to be absurd and persistent enough to be alarming. Over a year it left each spacecraft thousands of kilometres closer to the Sun than Newton and Einstein said it should be. For fourteen years the "Pioneer anomaly" was cited as possible evidence that gravity misbehaves at the edge of the solar system. Its eventual explanation involved no new physics, only decades-old blueprints, magnetic tapes rescued from disposal, and a model of how heat escapes a machine built in 1972.

The Facts

Pioneer 10 launched on 2 March 1972 and Pioneer 11 on 5 April 1973. Both were spin-stabilised, meaning they held attitude by rotating rather than by constantly firing thrusters. This made them unusually clean test bodies: the later Voyagers, which relied on frequent attitude burns, could not have detected an acceleration of 10⁻⁹ m/s² at all. Each Pioneer carried four SNAP-19 radioisotope thermoelectric generators (RTGs) fuelled by plutonium-238, mounted on booms away from the main body, with a large high-gain antenna pointed at Earth.

Tracking was done by the Deep Space Network using coherent two-way Doppler, with post-fit residuals at the millihertz level. Between roughly 20 and 70 AU the residuals showed a steady blue-shifted frequency drift of about 6 × 10⁻⁹ Hz/s. Interpreted as motion, that drift corresponded to a constant sunward acceleration. Anderson's 1998 paper gave ~8.5 × 10⁻⁸ cm/s²; the fuller 2002 study in Physical Review D, using Pioneer 10 data from 3 January 1987 to 22 July 1998 and Pioneer 11 data ending with the loss of coherent Doppler on 1 October 1990, refined it to a_P = (8.74 ± 1.33) × 10⁻¹⁰ m/s². That is the canonical number.

Pioneer 11's last telemetry was received on 11 February 1994 and its mission formally ended on 30 September 1995. Pioneer 10 returned its last telemetry on 27 April 2002 and its last faint carrier signal on 23 January 2003, at about 80 AU. A final listening attempt on 3-5 March 2006 heard nothing. Neither spacecraft can be re-examined.

Timeline

• 2 March 1972Pioneer 10 launches; Pioneer 11 follows on 5 April 1973.
• 1980A JPL analyst notices that both Pioneers are decelerating slightly more than the gravitational model predicts, with the craft beyond about 20 AU.
• 1 October 1990Pioneer 11 loses coherent Doppler capability at 31.7 AU; precision tracking of that spacecraft ends.
• 1994A formal JPL investigation of the anomaly begins, joined by Michael Martin Nieto of Los Alamos, who was struck by the numerical closeness of a_P to the product of the speed of light and the Hubble constant.
• 30 September 1995Last transmission from Pioneer 11.
• August 1998Anderson, Laing, Lau, Liu, Nieto and Turyshev publish the discovery paper in Physical Review Letters (81, 2858).
• 1999Jonathan Katz and Edward Murphy publish separate PRL comments arguing that RTG or electrical waste heat could produce the effect; Anderson's team disputes the magnitude.
• 11 April 2002Anderson et al. publish the 55-page Physical Review D study (65, 082004) fixing a_P = (8.74 ± 1.33) × 10⁻¹⁰ m/s² after an exhaustive systematics review. The authors say the most likely cause is an unknown systematic and admit they are split between gas leaks and heat.
• 23 January 2003Pioneer 10's last, telemetry-free signal is received at roughly 7.4 billion miles.
• 2003Louis Scheffer argues in Physical Review D that conventional forces, chiefly heat, can explain the whole anomaly.
• December 2005 to June 2006: Turyshev and colleagues report recovery of the full Doppler archive and telemetry files, much of it on obsolete 7- and 9-track tapes, with funding help from The Planetary Society.
• 2010Turyshev and Viktor Toth publish the comprehensive Living Reviews in Relativity survey; the anomaly is still described as unexplained.
• 2011Rievers and Lämmerzahl (Bremen) publish an independent thermal model in Annalen der Physik concluding thermal recoil "likely resolves" the Pioneer 10 acceleration.
• 19 August 2011Turyshev, Toth, Ellis and Markwardt (PRL 107, 081103) analyse the extended data set, 23.1 years for Pioneer 10 and 10.75 years for Pioneer 11, and find the anomaly is decaying with time.
• 12 June 2012Turyshev, Toth, Kinsella, Lee, Lok and Ellis publish "Support for the thermal origin of the Pioneer anomaly" (PRL 108, 241101).
• 2012Francisco, Bertolami, Gil and Páramos (Lisbon) publish a reflection-inclusive thermal model in Physics Letters B confirming the thermal result.
• July 2014Modenini and Tortora (Bologna) publish an independent orbit determination in Physical Review D concluding that no acceleration beyond the thermal one is needed.
• 2015Feldman and Anderson publish a cosmological "theory of inertial centers" alternative in IJMPD, while stating they remain sceptical of their own convention.

The Evidence

The core evidence was always the Doppler residual, and its most important property was one the early analyses could not see. With only 11.5 years of Pioneer 10 data, the anomaly looked constant. Constancy was the strongest argument against heat, because an RTG's output falls with the 87.74-year half-life of plutonium-238. When the recovered archive doubled the data span, the 2011 PRL analysis found a temporally decaying acceleration of about 2 × 10⁻¹¹ m/s² per year, improving the fit residuals by more than 10 percent over a constant model. It also found no evidence favouring a Sun-pointing direction over an Earth-pointing or spin-axis direction, which is exactly what an on-board force would look like, since the spin axis pointed at Earth.

The 2012 paper then supplied the mechanism. The JPL team, with the Applied Sciences Laboratory, built a finite-element thermal model of the spacecraft from blueprints and recovered documentation, since no CAD files existed. It used about 3,300 surface elements, 3,700 nodes and 8,700 linear conductors, and was driven by actual flight telemetry: power consumption of every subsystem and six on-board temperature sensors as boundary conditions. Modelled RTG thermal power came within 1 percent of the known value and RTG fin-root temperatures within ±2 K of telemetry.

Two heat sources dominated. RTG waste heat striking the back of the high-gain antenna was re-radiated forward, away from the Sun. Heat from electronics inside the equipment compartment also escaped mostly through the louvered rear face. Both push the spacecraft sunward. The paper parameterised the recoil as a_P = η_rtg(Q_rtg/mc) + η_elec(Q_elec/mc). The thermal model gave η_rtg = 0.0104 and η_elec = 0.406; an entirely independent fit to the Doppler data gave 0.0144 and 0.480. The thermal estimate amounted to roughly 80 percent of the Doppler-derived magnitude, about 35 percent from the RTGs and 45 percent from electrical heat. The 1σ error ellipses overlapped, and the authors concluded that no statistically significant anomalous acceleration remained.

Independent groups converged. Rievers and Lämmerzahl in Bremen, Francisco and colleagues in Lisbon, and Modenini and Tortora in Bologna each built their own thermal models; the Bologna team ran the full Doppler sets for both spacecraft through NASA's Orbit Determination Program with thermal recoil included and found the residual acceleration statistically compatible with zero for both Pioneers.

The Theories

What Doesn't Add Up

The 2012 authors themselves listed the open items, and they remain open. Pioneer 11 was never modelled at the same depth as Pioneer 10; the paper reports only a spot check. The anomalous spin-down of both spacecraft was left unaddressed, tentatively blamed on heat reflecting off instrument sunshades. The reported "onset" of the anomaly after Pioneer 11's Saturn encounter, with a lower and noisier value before it, was described as suggestive of solar-thermal mismodelling rather than proven to be so. Outgassing from surface materials was flagged as unexamined. The RTG paint properties could be measured in a thermal vacuum test of a hot RTG analogue; no such test has been reported.

There is also the matter of what the resolution rests on. The thermal model accounts for about 80 percent of the Doppler magnitude; full closure depends on the Doppler noise floor and the thermal error budget being large enough to overlap. That is a legitimate statistical conclusion, but it is not the same as measuring all of the force. A 2013 arXiv preprint (1307.0537) argued that the thermal papers were "open to questioning" while conceding a non-systematic anomaly was very unlikely. Feldman and Anderson's 2015 paper modelled the onset with a cosmological theory, assumed heat supplied only about 12 percent of the effect, and admitted the authors "remain skeptical" of their own clock-rate convention. Neither has gained traction.

Finally, the Pioneers cannot be re-observed. The evidence is closed; only the analysis can change.

Latest Developments

Nothing substantive has appeared in the last several years. Searches for 2023-2026 literature turned up no new peer-reviewed reanalysis of the Pioneer data, no revival of the anomaly by the MOND community, and no institutional statement reversing the 2012 conclusion. The 2010 Living Reviews survey, the 2012 PRL and the 2014 Bologna study remain the reference set.

What has changed is downstream. Turyshev and Toth showed in 2009 that the same techniques apply to New Horizons, predicting an RTG-driven acceleration of roughly 41 × 10⁻¹⁰ m/s², several times the Pioneer value, and warning it would degrade navigation accuracy. Thermal recoil is now a routine term in deep-space orbit determination rather than an anomaly. Cassini and Galileo, the other candidates for a comparison test, were deliberately destroyed in Saturn's and Jupiter's atmospheres.

Pioneer 10 is coasting toward Aldebaran, which it will approach in more than two million years; Pioneer 11 is headed toward Aquila. NASA's mission page, last updated in November 2024, records no change since the failed 2006 contact attempt. The case is closed by consensus, not by proof.

Sources

1. Indication, from Pioneer 10/11, Galileo, and Ulysses Data, of an Apparent Anomalous, Weak, Long-Range Acceleration — arXiv / Physical Review Letters, 1998
2. Study of the anomalous acceleration of Pioneer 10 and 11 — arXiv / Physical Review D, 2002
3. The Study of the Pioneer Anomaly: New Data and Objectives for New Investigation — arXiv, 2005
4. The Pioneer Anomaly in the Light of New Data — arXiv / Space Science Reviews, 2009
5. The Pioneer Anomaly — arXiv / Living Reviews in Relativity, 2010
6. High precision thermal modeling of complex systems with application to the flyby and Pioneer anomaly — arXiv / Annalen der Physik, 2011
7. Support for temporally varying behavior of the Pioneer anomaly from the extended Pioneer 10 and 11 Doppler data sets — arXiv / Physical Review Letters, 2011
8. Pioneer Anomaly is Fading — Physical Review Focus (APS), 2011
9. Modelling the reflective thermal contribution to the acceleration of the Pioneer spacecraft — arXiv / Physics Letters B, 2012
10. Support for the thermal origin of the Pioneer anomaly — arXiv / Physical Review Letters, 2012
11. Exotic explanation for Pioneer anomaly ruled out — Physics World, 2012
12. Pioneer Anomaly Solved! — The Planetary Society, 2012
13. ...and farewell to the Pioneer anomaly — Nature Physics, 2012
14. Pioneer 10 and 11 orbit determination analysis shows no discrepancy with Newton-Einstein's laws of gravity — arXiv / Physical Review D, 2014
15. Pioneer 10 Sends Last Signal — NASA Science, 2003 (updated 2024)

Northwest Flight 2501

A DC-4 with 58 aboard acknowledged a routine altitude denial over Lake Michigan and was never heard from again

The Hook

At 11:13 p.m. Central Standard Time on 23 June 1950, the captain of a Northwest Airlines DC-4 cruising at 3,500 feet near Benton Harbor, Michigan, asked Air Route Traffic Control for 2,500 feet. He gave no reason. Control refused because of other traffic. At 11:15 p.m. the flight acknowledged the refusal. That acknowledgment is the last thing anyone heard from Flight 2501. No distress call followed. About ten minutes later, by the Civil Aeronautics Board's estimate, the aircraft and the 58 people aboard went into Lake Michigan. The largest piece of wreckage recovered was, a Coast Guard captain said at the time, no bigger than a hand. Seventy-five years and roughly 700 square miles of side-scan sonar later, the airframe has never been found.

The Facts

The aircraft was a Douglas DC-4, registration N-95425, built in 1943 as a military C-54A, bought by Northwest in 1947 and converted to a 55-seat "cargo-coach" layout on 25 April 1950. The CAB report records no maintenance irregularities and a "mechanically okay" sign-off from the previous crew.

Flight 2501 was the daily New York to Seattle service via Minneapolis and Spokane. The crew were Captain Robert C. Lind, 35, of Hopkins, Minnesota, with Northwest since 1941 and 8,662 logged hours; First Officer Verne F. Wolfe, 35, of Minneapolis; and stewardess Bonnie Ann Feldman, 25. The 55 passengers, per the Michigan Shipwreck Research Association (MSRA), were 27 women, 22 men and six children.

The CAB gives departure from LaGuardia as approximately 1931, all times Central Standard. MSRA's page says 7:30 p.m. and van Heest's site 8:30 p.m. Eastern; the spread is time-zone confusion, and this file follows the CAB.

Weather dominated the flight from the start. A Northwest special forecast issued at 1545 warned of thunderstorms along a cold front with tops at 30,000 to 40,000 feet, moderate to severe turbulence, and "possible squall line development ahead of the front during the evening." Lind and the dispatcher chose 4,000 feet; ARTC assigned 6,000 because of traffic, then granted 4,000 over Cleveland at 2149. Forty minutes later ARTC asked for 3,500 feet to widen separation from an eastbound aircraft at 5,000 feet reporting severe turbulence over the lake. At 2251 Flight 2501 reported over Battle Creek at 3,500 feet, estimating Milwaukee at 2337.

At 2133, while the flight was airborne, the Weather Bureau issued a regional forecast describing a squall line across lower Michigan moving south. The CAB found this forecast "was not made available to the flight": it was not routinely broadcast, the Flight Advisory Weather Service did not ask ARTC to relay it, and Northwest's meteorologist doubted the squall line existed.

After the 2315 acknowledgment there was silence. Milwaukee radio declared the flight overdue at 2337; at 2358 ARTC alerted air-sea rescue; at 0530 the aircraft was presumed down, its fuel exhausted.

Timeline

• 23 June 1950, 1545 CST: Northwest issues a special thunderstorm forecast warning of possible squall line development.
• 23 June 1950, approx. 1931 CST: Flight 2501 departs LaGuardia with 55 passengers and three crew.
• 23 June 1950, 2133 CST: Weather Bureau forecast places a southward-moving squall line across lower Michigan; it is never passed to the flight.
• 23 June 1950, 2251 CST: Position report over Battle Creek at 3,500 feet.
• 23 June 1950, 2313 to 2315 CST: Near Benton Harbor, request for 2,500 feet denied and acknowledged. Last contact.
• 23 June 1950, approx. 2325 CST: CAB's estimated time of impact.
• 24 June 1950, 0530 CST: Aircraft presumed lost; daylight search begins on a fog-covered lake, initially off Milwaukee.
• 24 June 1950, 1830: Coast Guard cutter Woodbine finds an oil slick, debris and the aircraft logbook about 18 miles north-northwest of Benton Harbor.
• 25 June 1950Navy divers descend on sonar contacts in 150 feet of water and 30 to 40 feet of silt; nothing found.
• Late June 1950: Fragments and human remains wash ashore between Benton Harbor and South Haven; South Haven closes South Beach for nine days. The Korean War begins on 25 June and national attention moves on.
• 3 July 1950Riverview Cemetery, St. Joseph, records "body parts from airplane crash" in an unmarked plot.
• 13 to 14 July 1950: CAB public hearing in Chicago; more than 30 witnesses.
• 18 January 1951CAB adopts its report: probable cause cannot be determined.
• 2004MSRA and Clive Cussler's NUMA begin joint sonar searches out of South Haven.
• September 2008MSRA researcher Chriss Lyon finds the Riverview mass grave; a granite marker with all 58 names is unveiled 20 September 2008.
• 2013Cussler withdraws routine support; his sonar operator Ralph Wilbanks returns in 2015 to 2017 to chase MSRA leads.
• 24 June 2015Second unmarked mass grave at Lakeview Cemetery, South Haven, marked and dedicated on the 65th anniversary.
• 24 June 2025On the 75th anniversary MSRA announces it will search no longer.

The Evidence

The physical record is thin and entirely buoyant. The CAB lists everything recovered: a fuel tank float, foam rubber cushions, armrests, clothing, blankets, pillows, luggage, cabin lining, plywood flooring and other wooden parts. The cushions and armrests were "shredded from impact forces and cutting edges of the fuselage," which the Board read as proof of high-speed water impact. A plywood oxygen-bottle bracket indicated inertia forces acting forward, downward and to the left. Nothing showed any trace of fire.

Human remains were fragmentary. Coast Guard crews reported finding hands and ears; Berrien County coroner Louis Kerlikowske said it "must have been a terrific explosion to disintegrate the bodies so badly." Valerie van Heest, who has researched the case for two decades, told FOX 17 in 2025 that the largest piece found was a torso. Remains collected by the Coast Guard were cremated and buried at Riverview Cemetery; those gathered from South Haven beaches went into a second unmarked plot at Lakeview Cemetery, found in 2015 by sexton Mary Ann Frazier in an old plot register.

Witness testimony is vivid and unreliable. Retired Navy Lt. Cmdr. R. T. Helm told the South Haven Tribune he saw the aircraft pass low over his home and minutes later saw "a terrific flash out in the lake." William Bowie and a group at his gas station in Glenn, Michigan, described engines that "plunked," a yellow light trailing from a wing, and a flash "like the sun when it goes down." Both men testified at the CAB hearing. The Board did not adopt any of this as evidence of an explosion.

A fisherman snagged a twisted piece of light metal off South Haven on 12 July 1950; the CAB doubted it came from the DC-4. Sonar leads chased in 2015 to 2017 proved to be construction debris and scrap metal.

The meteorological evidence is the firmest. Pilots crossing southern Lake Michigan within an hour of the accident reported severe turbulence and constant lightning; several detoured south and three flights out of Detroit turned back. The CAB concluded the squall line's southern edge lay "at or near the location of the accident at time of accident."

The Theories

What Doesn't Add Up

The last-contact time is quoted as 11:13 or 11:15 p.m. Central (CAB), 12:13 a.m. Eastern (MSRA) and 1:13 a.m. "New York time" (New York Times, on daylight saving). They describe the same moment, but are routinely presented as a contradiction.

The reason for the descent request is a genuine conflict. The Times reported Lind asked to descend "because of a severe electrical storm." The CAB says no reason was given and no emergency declared.

The crash position is unexplained. A flight on Airway Red 57 from Glenn toward Milwaukee should not have been 18 miles north-northwest of Benton Harbor. Yet debris came ashore from Benton Harbor to Glenn, and Glenn residents insist the aircraft passed overhead. Whether it strayed, was blown off course, or the CAB's position is merely where an oil slick was found, is unresolved.

Whether remains came ashore is not in dispute among primary sources, though some retellings claim nothing was found. The Coast Guard record, the beach closure and two cemetery registers settle it.

Search figures drift between sources: nine, ten or eleven shipwrecks found; 450, 600 or 700 square miles covered; Cussler's involvement ending in 2013 or 2017. Any single number deserves caution.

The bodies. Coast Guard commanders in 1950 read the dismemberment as a mid-air explosion; the CAB read the same material as water impact. The airframe that would settle it is missing.

Latest Developments

The decisive recent event is the end of the search. On 24 June 2025, the 75th anniversary, MSRA executive director Valerie van Heest told the Detroit News the association would not go out again, after about 20 years, 10,000 hours and $500,000 of effort. "Part of me feels like we have failed," she said, adding that the work had kept the victims' memory at the forefront. The story ran nationally through the Associated Press, CBS and the New York Times. Van Heest's position is that the debris is fragmented below sonar resolution and buried in silt, and that only a future storm churning the lakebed is likely to reveal it. MSRA had continued annual sonar seasons through 2024 with its own deep-capable side-scan unit.

Discovery's Expedition Unknown special aired in February 2020, and Wikipedia records a segment on History's The Unbelievable with Dan Aykroyd in December 2023. Van Heest's "Fatal Crossing" exhibit is now at the Yankee Air Museum in Belleville; the Wisconsin Marine Historical Society reported in August 2025 that recovered personal effects are displayed at the Heritage Museum and Cultural Center in St. Joseph.

Nothing new has surfaced from the lake itself: no fresh sonar target, no wreckage, no revised official finding. The CAB's January 1951 verdict, that there is "not sufficient evidence upon which to make a determination of probable cause," remains the last word from any authority.

Sources

1. CAB Accident Investigation Report, File No. 1-0081, Docket SA-215 (full text) — Civil Aeronautics Board via Internet Archive, 1951
2. Northwest Airlines Flight 2501 (DC4) — Michigan Shipwreck Research Association, 2020
3. Northwest Airlines Flight 2501 — Valerie van Heest, 2020
4. Searchers end quest to find airliner that went missing over Lake Michigan 75 years ago — The Detroit News via The Spokesman-Review, 2025
5. Group ends search for missing airliner that crashed into Lake Michigan in 1950 — CBS News / Associated Press, 2025
6. The Disappearance of Flight 2501: Why the search is ending — FOX 17 West Michigan, 2025
7. The Search for Flight 2501 Has Ended — Wisconsin Marine Historical Society, 2025
8. Unsolved mystery of 75 years: The tragedy of Northwest Orient Airlines Flight 2501 — GlobalAir.com, 2025
9. Plane crash victims memorialized — The Herald-Palladium, 2008
10. New mass grave found for victims of NWA Flight 2501 — WZZM 13, 2015
11. The Mystery of Northwest Orient Flight 2501 — Fear of Landing (Sylvia Wrigley), 2015
12. Northwest Orient Airlines Flight 2501 — Wikipedia, 2025

Odd Radio Circles

Radio rings hundreds of kiloparsecs wide, centred on elliptical galaxies, invisible at every other wavelength

The Hook

In June 2020, a team led by Ray Norris of Western Sydney University and CSIRO posted a preprint with a deliberately plain title, "Unexpected Circular Radio Objects at High Galactic Latitude". It reported a class of objects found in pilot data from the Australian Square Kilometre Array Pathfinder (ASKAP), described them as "circular edge-brightened discs, about one arcmin diameter", and concluded that no mechanism examined "seems to be a compelling explanation".

The objects were rings of radio emission with no counterpart at any other wavelength. At the centre of two of them sat an ordinary-looking elliptical galaxy whose redshift, if the association was real, implied a ring roughly half a million parsecs across, ten to twenty times the diameter of the Milky Way. The team called them Odd Radio Circles, or ORCs, a name chosen to describe what was seen and nothing more. Six years on, the picture has grown more detailed and, in several respects, less tidy.

The Facts

The Evolutionary Map of the Universe (EMU) Pilot Survey observed roughly 270 square degrees of southern sky with ASKAP at 944 MHz. Three circular objects were found: ORC1 (ORC J2103−6200) and a close pair now labelled ORC2 and ORC3 (ORC J2058−5736). A fourth, ORC4 (ORC J1555+2726), was identified by Huib Intema in archival 325 MHz data from the Giant Metrewave Radio Telescope. ORC1 was confirmed with the Australia Telescope Compact Array at 2.1 GHz and the Murchison Widefield Array at 88–154 MHz, ruling out an imaging artefact.

ORC1 is about 80 arcseconds in diameter. Its central galaxy has a photometric redshift of 0.551, giving a physical diameter of about 520 kiloparsecs. Its integrated flux density is about 3.9 millijanskys at 1 GHz, a low-surface-brightness signal earlier surveys could not detect. The spectrum is steep, with a spectral index between roughly −1.2 and −1.6 depending on the frequency range fitted, characteristic of synchrotron emission from an ageing population of relativistic electrons rather than thermal gas.

A fifth object, ORC5 (ORC J0102−2450), was reported by Bärbel Koribalski and colleagues in 2021 from deep ASKAP observations of a field near the starburst galaxy NGC 253 taken between August 2019 and December 2020. Its ring is about 70 arcseconds, or 300 kiloparsecs at the central galaxy's redshift of about 0.27. The host is an elliptical with a stellar mass near 10^11 solar masses. Norris's 2022 paper put the probability of all three single ORCs having a radio-emitting galaxy at their centre by chance at around 10^-4. The central galaxies are therefore treated as hosts.

ORC2 and ORC3 lack a central galaxy and are treated as a separate class. By 2025 the definition adopted by Norris and colleagues required an edge-brightened radio circle, with no known emission at other wavelengths, surrounding a distant galaxy.

Timeline

• 2013The Giant Metrewave Radio Telescope records 325 MHz data later recognised as containing ORC4.
• 2019ASKAP's EMU Pilot Survey images the southern sky at 944 MHz; ORC1 and the ORC2/ORC3 pair are found. Observations of the NGC 253 field, which will yield ORC5, begin in August.
• June 2020Norris and colleagues post the discovery preprint (arXiv:2006.14805). On 6 June, MeerKAT observes ORC1 for 10 hours under Director's Discretionary Time.
• 2021The discovery paper appears in PASA (volume 38, e003). Koribalski and colleagues publish ORC5 in MNRAS Letters.
• April 2022Norris and colleagues publish the MeerKAT study of ORC1 in MNRAS, revealing internal arcs, polarisation and a tangential magnetic field.
• 2023Klaus Dolag and colleagues propose merger-driven shocks from cosmological simulations. MeerKAT observes the field containing ORC J0356−4216 on 12 and 15 October.
• 8 January 2024Alison Coil and colleagues publish Keck spectroscopy of ORC4 in Nature, arguing for a starburst-driven wind.
• 2024Esra Bulbul's team reports XMM-Newton detection of hot gas in the Cloverleaf system. Cristobal Bordiu's team reports Kýklos, an 80-arcsecond ring near the Galactic Centre. On 11 June, participants in a RAD@home online session identify a double-ring source in LOFAR data.
• 2025Norris and colleagues publish ORC J0219−0505, a 114-kiloparsec ring, from the MeerKAT MIGHTEE survey. Nikhel Gupta and colleagues report five new ORCs from the first year of the full EMU survey. On 25 September, Ananda Hota and colleagues publish RAD J131346.9+500320 in MNRAS. Sam Taziaux and colleagues publish polarimetry of ORC J0356−4216.
• 2026Anna Ivleva and colleagues publish a zoom-in merger simulation in A&A (February). Yiting Wang and Sebastian Heinz post a shock-bubble model (February). Miroslav Filipović and colleagues report the "Anglerfish" candidate (March). Maria Polletta, Coil and colleagues report J1248+4826, a 30-kiloparsec ring candidate (submitted 30 April).

The Evidence

The MeerKAT image of ORC1, at 6 arcseconds resolution, is about ten times deeper than the ASKAP discovery image. It resolves the outer ring into multiple arcs and shows radio knots inside, some coinciding with galaxies at photometric redshifts between 0.5 and 0.6 that agree with the host within one standard error, suggesting the ORC sits in a group. The polarisation map reveals a well-ordered magnetic field running tangentially around the ring, as expected from a shock compressing an ambient field. The interior arcs are less polarised than the outer ring, suggesting they are physically distinct structures.

Coil's Keck integral-field spectroscopy of ORC4 supplied the first non-radio data on an ORC host. It found strong [O II] emission across a region about 40 kiloparsecs in diameter at a spectroscopic redshift of 0.4512, an order of magnitude smaller than the radio ring but far larger than expected for a quiescent early-type galaxy. The gas shows a velocity gradient of about 200 kilometres per second, a velocity dispersion of about 180 kilometres per second and an equivalent width near 50 angstroms. The authors interpret this as shock-ionised gas falling back towards the galaxy after an outward-moving shock. The host's starburst ended roughly a billion years ago.

Taziaux and colleagues' study of ORC J0356−4216 combined MeerKAT at 1.28 GHz and ASKAP at 943 MHz. Two symmetric rings span about 2 arcminutes, or 668 kiloparsecs at the host's photometric redshift of 0.494. Both have steep spectral indices (−1.18 and −1.12), equipartition magnetic fields of about 1.7 to 1.8 microgauss, and fractional polarisation of 20 to 30 per cent with the field oriented along the rings.

RAD J131346.9+500320 is different in kind. In LOFAR's 144 MHz survey it shows two intersecting rings each about 300 kiloparsecs across, embedded in diffuse emission spanning about 800 kiloparsecs, around a host at photometric redshift 0.937. Its 144 MHz power of 2.27 × 10^26 watts per hertz is roughly two orders of magnitude above typical ORCs. The central source has a flat spectrum typical of an active radio core, and the field contains a group or poor cluster at the same redshift.

For the Cloverleaf, at redshift 0.046, XMM-Newton detected diffuse thermal X-ray gas over about 230 by 160 kiloparsecs, oriented perpendicular to the radio emission, with temperature components of 1.10 and 0.22 kiloelectronvolts. The total mass within R500 is about 2.6 × 10^13 solar masses, a low-mass group that optical data suggest is merging.

The Theories

What Doesn't Add Up

The class is heterogeneous, and the count depends on who is counting. Taziaux and colleagues in 2025 spoke of roughly ten securely identified ORCs; Polletta and colleagues in 2026 cited more than twenty confirmed plus a dozen candidates. The Cloverleaf is an ORC to Bulbul's team and, in Norris's words to Astronomy magazine, "interesting, but not an ORC", because it lacks the edge-brightening he regards as defining. Gupta's team added a further category, GLAREs, for 55 galaxies with diffuse ambient radio emission of uncertain relation to ORCs.

Sizes now span an uncomfortable range. ORC J0219−0505 is 114 kiloparsecs across; J1248+4826, if its group association holds, has a ring only about 30 kiloparsecs in radius with the host on its edge; RAD J131346.9+500320 spans 800 kiloparsecs with a radio power a hundred times the norm. A single mechanism producing all three would be remarkable.

Then there are the hosts. Every confirmed ORC with a central galaxy has a massive, old elliptical there. These galaxies do not generally show current jets, yet the AGN-relic models require powerful jets in the past, and the starburst model requires a merger-driven burst around a billion years ago. Both are plausible histories for a massive elliptical; neither is demonstrated for more than one object. And the simulations that get the shape right get the brightness wrong unless a reservoir of old relativistic electrons is assumed.

Latest Developments

The full EMU survey, which began in February 2023 and covered about 4,500 square degrees in its first year, is now the main engine of discovery. Gupta and colleagues' search of that year, combining object detection with visual inspection, yielded five new ORCs and two candidates. The MIGHTEE result showed that deeper surveys find smaller and fainter rings, implying ORCs are more common than the early estimate of one per 50 square degrees.

The most striking 2025 result came from outside the professional pipeline. RAD@home, an Indian citizen-science collaboratory founded in 2013 with no funding and no infrastructure, identified RAD J131346.9+500320 during a training session on 11 June 2024. It is the first ORC found by citizen scientists and the first found with LOFAR, at a redshift where the Universe was roughly half its present age.

Ivleva and colleagues' 2026 simulation gives merger shocks their most rigorous test to date and finds them necessary but not sufficient. Wang and Heinz's shock-bubble model is agnostic about redshift and host position, which would suit oddities like the off-centre host of J1248+4826. No spectroscopic study comparable to the Keck work on ORC4 has been published for a second ORC host; that remains the most obvious gap.

Sources

1. Unexpected Circular Radio Objects at High Galactic Latitude — arXiv / PASA, 2020–2021
2. Discovery of a new extragalactic circular radio source with ASKAP: ORC J0102–2450 — arXiv / MNRAS Letters, 2021
3. MeerKAT uncovers the physics of an Odd Radio Circle — arXiv / MNRAS, 2022
4. Ionized gas extends over 40 kpc in an odd radio circle host galaxy — Nature, 2024
5. Space Oddity: Uncovering the Origin of the Universe's Rare Radio Circles — UC San Diego Today, 2024
6. The Galaxy Group Merger Origin of the Cloverleaf Odd Radio Circle System — arXiv / A&A Letters, 2024
7. Mysterious Cloverleaf 'odd radio circle' could be a merger of a dozen galaxies — Astronomy, 2024
8. MeerKAT reveals a ghostly thermal radio ring towards the Galactic Centre — arXiv / A&A, 2024
9. MeerKAT discovery of a MIGHTEE Odd Radio Circle — MNRAS Letters, 2025
10. Discovery of Odd Radio Circles and Other Peculiars in the First Year of the EMU Survey using Object Detection — arXiv, 2025
11. RAD@home discovery of extragalactic radio rings and odd radio circles: clues to their origins — MNRAS, 2025
12. Most powerful 'odd radio circle' to date is discovered — Royal Astronomical Society via EurekAlert, 2025
13. Deep polarimetry study reveals double ring ORC-like structures — arXiv / A&A, 2025
14. Cosmological zoom-in simulation of odd radio circles as merger-driven shocks in galaxy groups — Astronomy & Astrophysics, 2026
15. Odd Radio Circles Modeled by Shock-Bubble Interactions — arXiv, 2026

Przybylski's Star

A faint star in Centaurus shows lines of elements that decay within decades; no one can prove they are real

The Hook

In 1961, Antoni Przybylski of Mount Stromlo Observatory published a one-page note in Nature under a title that now reads as an understatement: "A G0 star with high metal content." The object was HD 101065, an eighth-magnitude star in Centaurus, too faint for the naked eye. Its spectrum was so densely crowded with absorption lines that the familiar features used to classify stars were buried, and the strongest of the unfamiliar lines belonged to holmium, a rare earth so scarce that at the time it had not been securely identified in any other star, including the Sun.

That was the first anomaly. The second arrived in 2004, when a team led by Charles Cowley of the University of Michigan reported that statistical line matching indicated promethium and technetium in the same star. Neither element has a stable isotope. Promethium's longest-lived form decays with a half-life of 17.7 years. Four years later a Ukrainian-led group claimed absorption lines of nine actinides, from actinium to einsteinium, in a high-resolution spectrum taken with an 8.2-metre telescope of the European Southern Observatory. Einsteinium does not exist in nature on Earth; it was first identified in the debris of a 1952 thermonuclear test.

If those identifications are correct, something on or near the star is manufacturing radioactive elements faster than they can decay. If they are wrong, the error has survived two decades of scrutiny in a spectrum so complicated that more than half of its lines remain unidentified. Either answer is uncomfortable.

The Facts

HD 101065 is a main-sequence star with an effective temperature of roughly 6,400 K and a surface gravity of log g = 4.2, according to the model atmospheres of Shulyak and colleagues (2010). It is classified as a magnetic chemically peculiar star of the Ap type, and it is the prototype of the rapidly oscillating Ap (roAp) class. Asteroseismic modelling by Mkrtichian, Hatzes and Saio (2008), based on HARPS radial velocities taken on 3-6 March 2004, gave a best-fit mass of 1.50 solar masses, an age of about 1.6 billion years, and a polar magnetic field near 9 kilogauss. Its spectral lines are extremely sharp, with a projected rotational velocity of only 3.5 ± 0.5 km/s.

The chemistry is the core of the file. Lanthanides and other neutron-capture elements are overabundant by two to four orders of magnitude relative to the Sun, while iron and nickel are underabundant. Przybylski himself asked in a 1977 paper whether iron was present in the atmosphere at all. The rare earths are so dominant that they control the atmosphere's energy balance: Shulyak's group found that line blanketing by these elements heats the deep layers, cools the upper layers and shifts the emergent flux from the ultraviolet toward the infrared, so a standard solar-composition model of the star is simply wrong.

A longitudinal magnetic field of up to about -2.5 kilogauss was first measured in 1974 by Wolff and Hagen. In 2018 Swetlana Hubrig and collaborators, fitting 43 years of longitudinal field measurements with a dipole model, proposed a rotation period of approximately 188 years, while stressing that the solution was not unique. They measured a mean quadratic field of 1.6 ± 0.3 kilogauss. In 1978 Donald Kurtz, using the 20-inch telescope at the South African Astronomical Observatory, detected light variations of 10-20 millimagnitudes with a period of 12.15 minutes, the first rapid oscillations ever seen in an Ap star. Before that discovery, chemical peculiarity and pulsation were believed to be mutually exclusive.

The claimed radioactive elements are as follows. Cowley, Bidelman, Hubrig, Mathys and Bord (Astronomy & Astrophysics, 2004) reported Pm I and Pm II, Tc I and perhaps Tc II in HD 101065, and promethium in a second cool Ap star, HD 965. William Bidelman reported technetium and other unstable elements in a 2005 conference paper. Gopka, Yushchenko, Yushchenko, Panov and Kim (Kinematics and Physics of Celestial Bodies, 2008) identified lines of actinium (Z = 89), protactinium (91), neptunium (93), plutonium (94), americium (95), curium (96), berkelium (97), californium (98) and einsteinium (99) in a spectrum with resolving power R = 80,000.

Timeline

• 1913Antoni Przybylski is born in Rogoźno, Poland. He studies at the University of Poznań and works as a research assistant at its observatory.
• 1939-1941He serves as an artillery officer in the defence of Warsaw, is captured, and is held in a prisoner-of-war camp in Mecklenburg. In 1941 he escapes, crosses Germany largely by night, and is interned in Switzerland, where he earns a doctorate in technical sciences at the Zurich Polytechnic in 1949.
• 1950He emigrates to Australia and is digging cable trenches under a two-year labour bond when a government officer forwards his credentials to Richard Woolley, director of Mount Stromlo. Colleagues rename him "Bill".
• 1954Przybylski receives the first PhD ever awarded by the Australian National University, for work on stellar atmospheres.
• 1961His Nature note announces HD 101065 as a G0 star of unusually high metal content. Kron and Gordon, and Przybylski himself, note the extraordinary line blanketing.
• 1963Przybylski reports lines of holmium in the star (Acta Astronomica); with Kennedy he publishes a fuller description of the spectrum.
• 1974Wegner and Petford publish the first detailed abundance analysis; Wolff and Hagen measure a strong magnetic field.
• 1975Przybylski asks in print whether the peculiarities could be explained by fission of transuranium elements.
• 1978Kurtz discovers the 12.15-minute oscillations, founding the roAp class.
• 1980Whitmire and Wright argue in Icarus that a civilisation disposing of fissile waste in its star would leave a detectable spectral signature, most plausibly in A5-F2 stars.
• 21 September 1985Przybylski dies suddenly in Queanbeyan, New South Wales, after 28 years on the Mount Stromlo staff. (The Canberra Museum and Gallery's Stromlo exhibition gives 1986; the contemporaneous ANU obituary gives 1985.)
• 2000Cowley, Ryabchikova, Kupka, Bord, Mathys and Bidelman publish a comprehensive abundance study in MNRAS.
• 2004Cowley and colleagues report promethium and technetium. Gopka's group presents "radioactive shells" in peculiar stars at IAU Symposium 224.
• 2007Fivet and colleagues compute oscillator strengths for 46 Pm II lines and note that the detections are not entirely certain.
• 2008Gopka and colleagues publish the actinide identifications; Gopka, Ul'yanov and Andrievsky propose a neutron-star companion; Mkrtichian, Hatzes and Saio publish the HARPS asteroseismology; Tatiana Ryabchikova argues that lines attributed to promethium are reproduced by cerium.
• 14 March 2017Dzuba, Flambaum and Webb post a preprint suggesting the actinides are decay products of a metastable superheavy nucleus with 184 neutrons. It appears in Physical Review A on 30 June 2017.
• 2018Hubrig and colleagues propose the tentative 188-year rotation period.
• 2022-2023Andrievsky revives the neutron-star model as a gamma-ray pulsar; Andrievsky and Kovtyukh find no deuterium; two further papers conclude that neither promethium nor technetium can be confirmed.
• May 2026Centauri Dreams reports that a second peculiar star, HD 25354 in Perseus, is now under investigation for unstable elements.

The Evidence

The strongest evidence is also the oldest. The rare-earth overabundance is not in dispute; it has been measured repeatedly since 1974 and is large enough to reshape the atmosphere. The sharp lines, the kilogauss field and the 12-minute pulsations are secure.

The evidence for radioactive elements rests on wavelength coincidences. Cowley's 2004 paper used both traditional identification and wavelength coincidence statistics, and the authors wrote that the evidence would justify declaring promethium present "without hesitation" if any isotope were stable. They discussed two escape routes: chance coincidences in a spectrum with a line every few hundredths of a nanometre, and contamination of the laboratory sources used to measure promethium wavelengths. Gopka's 2008 actinide paper used the same approach on a better spectrum, and reported all elements from Z = 84 to 99 except astatine and francium.

The counter-evidence is more recent and more quantitative. Ryabchikova pointed out in 2008 that two lines at 665.907 and 677.228 nm attributed to promethium are well reproduced by Ce II, one of the most overabundant species in the star. In 2023 Andrievsky, Korotin, Werner and Kovtyukh selected the three least-blended promethium lines and synthesised the surrounding spectrum with a stratified model atmosphere. The Pm II line at 557.602 nm, considered the most reliable, sits inside a blend of roughly thirty Co I hyperfine components plus a Fe I line at 557.609 nm and possibly Nd II at 557.599 nm. Fitting it gave a promethium abundance of log ε(Pm/H) = +1.8, but applying that abundance to the two Pm I lines produced results inconsistent by more than two orders of magnitude. The authors concluded that "at present it is impossible" to claim promethium in the star. A companion paper found that the Tc I resonance line at 429.706 nm is not visible and reduced technetium to an upper limit of log ε(Tc/H) = 2.5.

No one has yet published an equivalent line-by-line synthesis of the nine actinides.

The Theories

What Doesn't Add Up

The star's basic parameters disagree with each other. The asteroseismic model of 2008 implies a polar field near 9 kilogauss; the direct magnetic measurements give a mean quadratic field around 1.6 kilogauss, and the 2023 Odessa papers quote 500-1,500 gauss from the same 2018 study. The 188-year rotation period rests on 23 per cent phase coverage.

Abundances derived from different ionisation stages of the same element differ by one to two orders of magnitude, which Shulyak's group attributed to unmodelled stratification. That same problem undermines any claim about a trace element, and equally any refutation.

The Gopka identifications were made on a spectrum from an 8.2-metre telescope at R = 80,000, better data than most claims of exotic detections rest on, yet the paper was published in a journal with limited circulation and has attracted no independent replication in eighteen years. David Kipping and Jason Wright have both remarked on the lack of follow-up. Charles Cowley, who spent six decades on the star, died in 2024 without the question resolved.

Latest Developments

The most significant work of the 2020s is subtractive. Between 2022 and 2023 Sergei Andrievsky and colleagues in Odessa and Tübingen, working from archival HARPS and VLT spectra, published three papers that respectively failed to find deuterium (upper limit [D/H] below -4), failed to confirm promethium, and reduced technetium to an upper limit. The same group continues to advocate a gamma-ray pulsar companion, so these are not hostile analyses; they are the model's own authors reporting that the direct evidence for ongoing neutron irradiation is absent.

No detection of a superheavy parent nucleus in HD 101065 has been reported, and the island of stability remains unconfirmed in the laboratory.

In May 2026 Centauri Dreams, prompted by a technosignature discussion on the Event Horizon podcast, revisited the star and noted that HD 25354 in Perseus, where Jaschek and Brandi reported transuranium elements in 1972, is again being examined. As of this writing, no new high-resolution actinide analysis of Przybylski's Star has appeared in the refereed literature.

The case therefore stands where it did in 2008, with one difference. The two lightest radioactive claims have been tested with modern synthesis and did not survive. The nine heaviest have not been tested at all.

Sources

1. On the possible presence of promethium in the spectra of HD 101065 (Przybylski's star) and HD 965 — Astronomy & Astrophysics, 2004
2. Identification of absorption lines of short half-life actinides in the spectrum of Przybylski's star (HD 101065) — Kinematics and Physics of Celestial Bodies, 2008
3. Isotope shift and search for metastable superheavy elements in astrophysical data — arXiv / Physical Review A, 2017
4. An enigma of Przybylski's star: is there promethium on its surface? — arXiv / Astronomische Nachrichten, 2023
5. Abundance of radioactive technetium in Przybylski's star revisited — arXiv, 2023
6. Probing the Przybylski star for deuterium — arXiv / Astronomische Nachrichten, 2023
7. Magnetic and pulsational variability of Przybylski's star (HD 101065) — arXiv / MNRAS, 2018
8. The detection of the rich p-mode spectrum and asteroseismology of Przybylski's star — Astronomy & Astrophysics via ADS, 2008
9. Nuclear waste spectrum as evidence of technological extraterrestrial civilizations — Icarus via ADS, 1980
10. On the radioactive shells in peculiar main sequence stars: the phenomenon of Przybylski's star — Proceedings of the IAU, 2004
11. Obituary: Antoni (Bill) Przybylski (1913-1985) — Obituaries Australia, Australian National University, 1985
12. Antoni Przybylski (1913-1986) — Canberra Museum and Gallery, Outer Space: Stromlo to the Stars
13. Przybylski's Star III: Neutron Stars, Unbinilium, and aliens — AstroWright (Jason Wright, Penn State), 2017
14. Przybylski's Star: Still Bizarre After All These Years — Centauri Dreams, 2026
15. Rapidly oscillating Ap star — Wikipedia (citing Kurtz 1978, IBVS 1436)

The SS Waratah

A 9,339-ton liner, 211 people, one exchange of lamp signals off the Bashee River, then nothing

The Hook

At about 06:00 on 27 July 1909, the fourth officer of the Clan Line steamer Clan MacIntyre worked a signal lamp at a larger, faster ship that had come up astern in the night off the mouth of the Bashee River. The exchange was routine: name, port of departure, destination. The other ship identified herself as the Waratah, out of Durban for Cape Town and London. She drew ahead, and by about 09:30 she had gone over the horizon.

By that evening the weather on the Eastern Cape coast had broken into a storm that the master of the Clan MacIntyre would later describe as the worst he had seen in thirteen years at sea. The Waratah was due in Cape Town on 29 July. She did not arrive that day, or any day after it. No lifeboat, no body and no piece of identifiable wreckage has ever been recovered from a ship of more than 9,000 tons that vanished within sight of a well-travelled coast. The Board of Trade, sitting in London eighteen months later, could say only that she had probably capsized, and that "what particular chain of circumstances brought about this result must remain undetermined."

The Facts

The Waratah was ordered in September 1907 by W. Lund and Sons for their Blue Anchor Line service between London and Australia, and built by Barclay, Curle and Company at Whiteinch, Glasgow, as yard number 472. Lloyd's Register records her at 9,339 gross register tons, 465 ft (141.7 m) between perpendiculars and 59 ft 4 in (18.08 m) in beam, with twin quadruple-expansion engines driving two screws. Some later accounts give her length as 500 ft; that figure does not match the registered dimensions. She had a cellular double bottom, eight watertight compartments, lifeboat and raft capacity for 921 people, and no wireless, which was not unusual in 1908.

She was a mixed passenger and cargo ship, designed to carry emigrants outward and to return laden with frozen meat, wool, dairy produce and metal ore. Her design was based on the company's earlier Geelong, which her master, Captain Ilbery, had previously commanded. Most sources give his name as Josiah Edward Ilbery; the Company of Master Mariners of Australia and Discovery UK render it Joshua. He was the line's senior master, with roughly three decades of experience.

Her maiden voyage left London on 5 November 1908 and returned on 7 March 1909. During it a fire broke out in a starboard bunker adjoining the engine room and burned intermittently from 6 to 10 December. The second voyage left London on 27 April 1909, loaded lead concentrate at Adelaide and wool, flour and frozen produce at Melbourne and Sydney, and sailed from Adelaide on the homeward leg on 7 July. Cargo figures conflict: Wikipedia, drawing on Harris, puts the lead concentrate at around 1,000 tons; the Master Mariners of Australia give 300 tons; the State Library of South Australia, citing Australian manifests, puts total cargo at between 10,000 and 10,987 tons, far more than the owners publicly claimed.

She arrived at Durban on 25 July 1909, a day early. Several passengers disembarked there as planned. One, Claude Gustav Sawyer, an engineer of about 54 booked to Cape Town with an option to continue to London, left the ship although he had intended to go on. She sailed at about 20:15 on 26 July with 211 people aboard: 92 passengers and 119 crew.

Timeline

• 12 September 1908Waratah launched at Barclay Curle's Whiteinch yard; sea trials 23 October.
• 5 November 1908Maiden voyage departs London; bunker fire near the engine room 6-10 December.
• 27 April 1909Second voyage departs London; Adelaide 6 June; loading at Melbourne and Sydney.
• 7 July 1909Departs Adelaide for Durban, Cape Town and London.
• 25 July 1909Arrives Durban. Claude Sawyer disembarks.
• 26 July 1909, about 20:15: Sails from Durban with 211 aboard.
• 27 July 1909, about 06:00-09:30: Signals and overtakes the Clan MacIntyre off the Bashee River. Last confirmed sighting.
• 27 July 1909, evening: The Harlow sees smoke and two flashes astern; the Guelph reads only "T-A-H" from a passing ship.
• 27-28 July 1909: Severe storm on the Eastern Cape coast.
• 29 July 1909Waratah fails to arrive at Cape Town.
• 1 August 1909The tug T. E. Fuller begins the search; cruisers Pandora, Forte and Hermes follow.
• 4 August 1909Sawyer cables his wife, "Thought Waratah top heavy. Landed Durban."
• September-November 1909: The chartered Sabine searches some 14,000 nautical miles of the southern Indian Ocean without result.
• 15 December 1909Waratah posted missing at Lloyd's.
• Early 1910The Wakefield, funded by Australian relatives, searches for four months, landing men on the Kerguelen islands.
• 15 December 1910Board of Trade investigation opens at Caxton Hall, Westminster; findings delivered February 1911.
• 1929Edward Joe Conquer first publicly describes watching a ship founder off the Xhora River on 28 July 1909.
• 1983-1997Emlyn Brown leads repeated searches, from 1987 with Clive Cussler's NUMA.
• 14 July 1999Brown announces a wreck he identifies as the Waratah.
• January 2001Submersible footage shows it is the Nailsea Meadow, torpedoed in the Second World War.
• 2004Brown abandons the search.
• 27 July 2009Centenary plaque unveiled at the Queenscliffe Maritime Museum, Victoria.
• 4 February 2025D. J. Pons publishes a navigational reconstruction proposing sinking positions off Algoa Bay.

The Evidence

The confirmed record ends with the Clan MacIntyre. Her fourth officer, W. E. Carson, initiated the signals; the ships exchanged names and destinations; the Waratah, making about 13 knots, drew away and was last seen at about 09:30. Her master testified at the inquiry to the violence of the weather that followed.

Three further sightings were reported, none confirmed. At about 17:30 on 27 July the steamer Harlow, roughly 180 nautical miles from Durban, saw heavy smoke from a ship astern, then after dark her running lights some 10 to 12 nautical miles off, then two bright flashes, after which the lights were gone. The Harlow's master thought of explosions; her mate thought of bush fires ashore, common on that coast in July. The master accepted the mate's view, made no log entry, and reconsidered only after news of the disappearance. At about 21:30 the Union-Castle liner Guelph, northbound for Durban, exchanged lamp signals with a passing ship but could read only the last three letters of her name as "T-A-H". The inquiry noted that these sightings could not all be of the Waratah unless she had reversed course after passing the Bashee River.

The fourth account was not before the court. Edward Joe Conquer, a Cape Mounted Rifleman on exercises at the Xhora River mouth on 28 July, later said that through a telescope he watched a steamship matching the Waratah's description labouring south-west, roll heavily to starboard and disappear under a following sea. He reported it to his sergeant, who did not take it up, and did not go public until 1929. In 1925 a South African Air Force pilot, D. J. Roos, reported a wreck seen from the air off the Transkei coast; it was never located again.

Nothing tangible has been recovered. Reports of bodies off the Bashee River in August 1909 were investigated by the tug Harry Escombe, which found whale offal and dead skate. A life preserver said to bear the ship's name washed up in New Zealand in 1912, but could equally have been lost overboard in Australian waters. Wreckage reported at Mossel Bay in 1910 and near East London in 1939 was never tied to the ship.

The Board of Trade inquiry sat at Caxton Hall from 15 December 1910 into February 1911. The expert evidence was that the ship was properly designed and built and classed +100 A1 by Lloyd's. Against this stood passengers and crew from both voyages who described a persistent list, a slow return from a roll, and a tendency to bury her bows in a head sea. Professor William Bragg, who sailed on the maiden voyage, described a list to starboard held for hours and concluded that her metacentre sat just below her centre of gravity. For every such witness the court heard another who found her comfortable and steady. The builders produced calculations showing she was not top-heavy even with the deck coal several witnesses said she carried out of Durban.

The court found "that the ship was lost in the gale of the 28th of July, 1909, which was of exceptional violence for those waters and was the first great storm she had encountered," and that it "on the whole, inclines to the opinion that she capsized." It found no fault with construction, equipment or stowage, but recommended that a committee of experts establish minimum stability requirements. It also observed that the company's correspondence with Captain Ilbery after the maiden voyage covered fittings, cabins and ventilation and never raised the ship's handling, and that the company had never asked. The Blue Anchor Line, under-insured and publicly damaged, sold its fleet to P&O and went into voluntary liquidation in 1910.

The Theories

What Doesn't Add Up

The sightings cannot all be true. If the Waratah passed the Clan MacIntyre off the Bashee River at 06:00 and was still steaming south-west, she could not have been the ship the Harlow saw astern that evening, nearer Durban, unless she had turned back into the storm. The court said as much. Conquer's later account places a foundering ship off the Xhora River on 28 July, the day after the Harlow's flashes.

The stability evidence is irreconcilable. Bragg, a future Nobel laureate, and Sawyer, an engineer, described behaviour consistent with a negative metacentric height; officers, stokers and other passengers described an easy, comfortable ship. Both may have been right: a slightly tender liner would feel pleasant to a lay passenger and alarming to an expert, and would depend on correct loading that the company never seems to have discussed with her master.

The cargo figures do not agree. The owners said the ship was lightly laden; the Australian manifests examined by the State Library of South Australia indicate she was carrying over 10,000 tons.

And the false find of 1999 is itself a puzzle. Brown's side-scan sonar showed a wreck of the right size in 117 metres of water between the Xora and Bashee rivers, exactly where Conquer's sighting placed it. It was the Nailsea Meadow, a 420 ft freighter torpedoed by U-196 in May 1943 (NUMA's contemporaneous note gives 1942; Lloyd's Register-based records give 1943). A search that finds a ship of the right dimensions in the right place, and is still wrong, is a caution against every confident location proposed since.

Latest Developments

The last major search effort ended in 2004 when Emlyn Brown, after 22 years and expeditions in 1983, 1989, 1991, 1995, 1997 and 1999-2001, said he had exhausted all options and no longer knew where to look. In 2009, the year of the centenary, he offered the view that the ship had capsized and drifted off the shelf before sinking; a centenary plaque was unveiled at the Queenscliffe Maritime Museum in Victoria on 27 July 2009.

The most substantive recent work is academic. In February 2025 the Journal of Maritime Archaeology published an open-access paper by D. J. Pons that reconstructs the Waratah's course from the Clan MacIntyre's account, models the meteorological front and the Agulhas Current, and concludes that the ship met the storm at the shelf edge off Algoa Bay, near Gqeberha (Port Elizabeth). The paper argues that the weather north of East London was relatively insignificant, weakening the case for a loss near the Bashee River where nearly all searching has concentrated, and identifies two candidate positions at the shelf break near submarine canyon heads, at roughly 34.35°S 26.26°E and 34.39°S 25.94°E. No expedition has yet tested them.

No new survey, wreck identification or official reopening of the case was found in the period 2020 to September 2026. The Waratah remains posted missing at Lloyd's, as she has been since 15 December 1909.

Sources

1. SS Waratah — Wikipedia, accessed 2026
2. Seascape Factors in the Loss of the SS Waratah — Journal of Maritime Archaeology (Springer), 2025
3. Waratah: a fresh look at the legend of the bloody swordsman — A Fortean in the Archives (transcribing The Times, 17 December 1910), 2009
4. Michael Cheika and the SS Waratah — The Roar (quoting the Board of Trade findings), 2016
5. 'Australia's Titanic' - the sinking of the SS Waratah — State Library of South Australia, accessed 2026
6. 1909 Last sighting of steamer SS Waratah — Company of Master Mariners of Australia, accessed 2026
7. Waratah — National Underwater and Marine Agency, 2001
8. Search for Waratah goes on after 'false' find — IOL / Sapa, 2001
9. Discovery of the wreck of passenger liner SS Waratah — South African History Online, 2011
10. South Africa: Waratah riddle solved at last, wreck found off Transkei coast — allAfrica, 1999
11. SS Nailsea Meadow — Wikipedia, accessed 2026
12. What Happened to the SS Waratah and Has it Been Found? — Discovery UK, 2024

The Yuba County Five

Five friends left a basketball game in Chico; their car was found 70 miles away on a snowbound mountain road

The Hook

The car was not broken. When deputies hot-wired the turquoise-and-white 1969 Mercury Montego on 28 February 1978, the engine started immediately. The tank was a quarter full. The snow it sat in was shallow enough that five grown men could have pushed it clear. Inside were snack wrappers, basketball programmes and neatly folded road maps. Outside, at 4,400 feet on an unploughed Forest Service road in the Plumas National Forest, there was nothing: no footprints that survived the storm, no bodies, no explanation.

The five men who had ridden in that car were 70 miles from where they had last been seen and nowhere near the road home. Four of them would be found in June, once the snow melted. One of them, Ted Weiher, had survived for weeks in a locked Forest Service trailer stocked with food, fuel and clothing, and had died anyway. The fifth, Gary Mathias, has been missing for 48 years.

The Facts

The five were Theodore "Ted" Weiher, 32; Jack "Doc" Madruga, 30; William "Bill" Sterling, 29; Gary Mathias, 25; and Jackie Huett, 24. All lived with their parents in Marysville or Yuba City. Weiher, Huett and Sterling had intellectual disabilities; Madruga was undiagnosed but described by his family as slow. Mathias, an Army veteran with a psychiatric discharge, had been diagnosed with paranoid schizophrenia. By 1978 he was on regular medication and working in his stepfather's gardening business; his stepfather, Robert Klopf, and a local doctor told the Washington Post he had been stable for two years.

The men played together on the Gateway Gators, a basketball team run through Gateway Projects, a Yuba City vocational rehabilitation programme. On the morning of 25 February they were due to play in a Special Olympics tournament; the Appeal-Democrat reports it was in Rocklin, while Den of Geek places it in Sacramento. The winners would earn a trip to Los Angeles. Their uniforms were laid out at home before they left.

On the evening of Friday 24 February, Madruga drove the group roughly 50 miles north to California State University, Chico, to watch UC Davis play Chico State. After the game they stopped at Behr's Market, three blocks from campus, shortly before its 10 p.m. closing, and bought pies, candy bars, two Pepsis and milk. The clerk remembered them because the group delayed her closing. That was the last confirmed sighting of all five.

When they had not returned by morning, the families began calling one another. Madruga's mother tried to report her son missing at around 8 a.m. on 25 February and, according to the Yuba County Five podcast and family accounts, was told to wait; a formal report was filed that evening.

On Tuesday 28 February, a Forest Service employee who had seen the Montego on the Oroville-Quincy road on 25 February recognised it from a bulletin and led deputies to it, near Rogers Cow Camp in Butte County. A snowstorm hit the same day. Two days later searchers in snowcats nearly became lost themselves, and the ground search was suspended. It was formally called off on 30 March.

On 4 June a group of motorcyclists found a broken window on a Forest Service trailer at the Daniel Zink Campground and, inside, Weiher's body. Over the following days the remains of Madruga, Sterling and Huett were recovered from the road corridor between the car and the trailer. Mathias was not found. The three counties involved ended the search for him on 20 June 1978.

Timeline

• 23 February 1978A Forest Service snowcat runs the road to the Daniel Zink trailers to clear snow, leaving a packed track.
• 24 February 1978, evening: Joseph Schons (spelling varies) becomes stuck near the snow line on the same road and suffers a heart attack.
• 24 February 1978, about 10 p.m.: The five men buy snacks at Behr's Market in Chico and leave. Last confirmed sighting.
• 24-25 February 1978, around midnight: Schons reports seeing headlights, a group of people, later flashlights, near where the Montego was afterwards found.
• 25 February 1978Families report the men missing. A Forest Service employee sees the Montego but does not report it until 28 February.
• 28 February 1978The Montego is located on the Oroville-Quincy road. A storm drops nine inches of snow on the upper mountain.
• 2 March 1978Snowcat teams nearly lose searchers in drifts; the mountain search is halted, and formally suspended on 30 March.
• 3 March 1978A shop worker in Brownsville tells deputies four of the men stopped at her store in a red pickup on 25 February.
• 4 June 1978Weiher's body is found in the Forest Service trailer.
• 5-6 June 1978: Weiher is identified; remains of Madruga and Sterling are found on opposite sides of the road, 11.4 miles from the car (Washington Post).
• 7-8 June 1978: Huett's father finds his son's remains near the trailer; a skull is recovered the following day and matched by dental records.
• 20 June 1978Yuba, Butte and Plumas counties end the search for Mathias.
• 2019 or 2020 (date disputed): A Sheriff-Coroner memo states Mathias is believed to be a victim of foul play; the case remains open as a missing person/homicide matter.
• October 2023The memo is released through a public records request, according to ABC News (Australia).
• 3 April 2024Netflix releases Files of the Unexplained, including an episode on the case.
• 4 August 2025Action News Now reports the case remains open; the Sheriff's Office asks for tips at 530-749-7777.

The Evidence

The car. Investigators found no dents, gouges or heavy mud on the Montego's undercarriage, including its low-hanging muffler, despite the rutted road. The Washington Post reported the police view that the driver either used exceptional care or knew the road. The keys were never recovered. The car was unlocked and a window was down. Madruga's family said he never let anyone else drive.

The witness. A 55-year-old Sacramento man told police he became stuck about 50 yards above where the Montego was later found, had a heart attack, and spent the night in his running car. Around midnight he saw headlights and what looked like a group of people, possibly including a woman with a baby; he called out and the lights went off. Later he saw flashlights that also went out when he shouted. When his car ran out of fuel he walked roughly eight miles down to the Mountain House lodge, passing the empty Montego. Doctors confirmed a mild heart attack. He later told the Los Angeles Times he had been half-conscious and hallucinating and was unsure whether the second vehicle was real. His surname is spelled Schons in the Los Angeles Times, the Yuba County Five podcast and Veriarch, and Shones in the Washington Post and ABC News (Australia); this file uses Schons.

The trailer. Weiher lay on a bed under eight sheets pulled up over his head. His feet were severely frostbitten. He had lost between 80 and 100 pounds of his roughly 200. Beard growth put his survival at eight to thirteen weeks. On a table were his wallet with cash, his ring, his necklace, and a gold Waltham watch, missing its crystal, that no family recognised. His leather shoes were missing. Mathias's tennis shoes were in the trailer. Some three dozen C-ration cans from an outside shed had been opened, at least one with a P-38 military can opener, which of the five only Madruga and Mathias would likely have known. A locker of dehydrated food sufficient for a year was untouched. A propane tank that would have fed the trailer's heater was never opened. No fire was lit, though matches, paperback books and wooden furniture were present. Heavy forestry clothing remained in storage.

The road. Madruga's remains were partially consumed by animals; his hand was around his watch. Only scattered bones remained of Sterling. Huett's backbone, jeans and shoes were found under manzanita northeast of the trailer; Wikipedia gives two miles, the podcast timeline less than one. A quarter mile northwest of the trailer, three Forest Service blankets and a rusted, switched-off flashlight lay beside the road.

Distances vary: the Washington Post gives 19.4 road miles from car to trailer; the Appeal-Democrat and the podcast give about 12, apparently straight-line; Action News Now and A&E round to 20.

The Theories

What Doesn't Add Up

The drive itself. Five men who had prepared for a game the next morning drove 70 miles in the wrong direction, up a road none of them knew, in light clothing, and did it without scraping the car.

The abandonment. A working car with fuel was left with keys gone and a window down. Whoever left it walked uphill toward nothing rather than downhill toward the lodge Schons reached on foot.

The trailer. A man starved for weeks beside a year's worth of food and a propane heater that needed only a valve turned. Someone tucked eight sheets over him; his frostbitten, near-gangrenous feet would likely have made it too painful to do himself, according to the Los Angeles Times. Someone wearing his shoes walked out and vanished.

The memo. The Sheriff-Coroner's document is dated 15 November 2019 in the text reproduced by a 2024 change.org petition, and 8 October 2020 by Veriarch; ABC News (Australia) describes it as mid-October 2020. All versions agree on the wording and on an instruction that the letter not be forwarded to the Mathias family. The Sheriff's Office has not publicly explained what prompted it.

The trailer report. Two Forest Service workers said on 9 June 1978 that they had told Butte County about the trailers early in the search. If so, the trailer was not checked while at least one man was alive inside it.

Latest Developments

The case remains open and unsolved. The most significant recent change is the October 2023 release, via records request, of the Sheriff-Coroner's memo classifying Mathias's disappearance as a missing person/homicide matter. Tony Wright's book Things Aren't Right (January 2024) and Netflix's Files of the Unexplained (April 2024) brought renewed attention; a December 2024 petition called for an independent review and modern forensic testing. In February 2025 KCRA marked the 47th anniversary, and in August 2025 Action News Now reported the Sheriff's Office still seeking tips. No remains have been identified as Mathias, no DNA result has been announced, and no suspect has been named. No confirmed report of the death of Mathias's mother, Ida Klopf, was located for this file.

Sources

1. 5 'Boys' Who Never Came Back — The Washington Post, reprinted at dyatlovpass.com, 1978
2. Out in the Cold — The Sacramento Bee, reprinted at dyatlovpass.com, 2019
3. Five men disappear on a snowy February night — ABC News (Australia), 2024
4. New book examines the 'Yuba County Five' — Appeal-Democrat, 2024
5. Unsolved Mystery: What happened to Gary Mathias and the 'Yuba County 5' — Action News Now, 2025
6. 47 years later, the disappearance of the Yuba County Five remains an unsolved mystery — KCRA via CNN Newsource/KVIA, 2025
7. What Happened to the Yuba County Five? — A&E, 2025
8. Files of the Unexplained: What Netflix's Yuba County Five Doc Leaves Out — Den of Geek, 2024
9. About the Case — Yuba County Five podcast (Mopac Audio), 2022
10. The Yuba County Five: A Case Reclassified — Veriarch, 2025
11. Demand Justice: Reevaluate the Yuba County Five Case — Change.org petition, 2024
12. Yuba County Five — Wikipedia, 2026

The Wow! Signal

A 72-second narrowband burst at the hydrogen line, seen once by one horn of the Big Ear and never again

The Hook

The evidence fits on one line of fan-fold computer paper. In channel 2 of a 50-channel printout from the Ohio State University Radio Observatory, six characters appear in sequence: 6EQUJ5. Around them is a red circle, and in the margin, in the same red pen, a single word: "Wow!"

The annotation was made by Jerry R. Ehman, a volunteer astronomer reviewing printouts at home a few days after the telescope recorded them. The instrument, a drift-scan telescope known as the Big Ear, had been running the world's first full-time search for extraterrestrial radio signals since 1973. Nothing it recorded before or after resembled that line.

The signal was strong, narrower than 10 kHz, and close to the 1420 MHz line of neutral hydrogen, the frequency SETI theorists had long argued an alien transmitter would choose. It rose and fell exactly as a fixed point in the sky should as the Earth's rotation carried it through the beam. Then it failed to reappear in the second horn three minutes later, as a real source should have. It has not been detected since.

The Facts

The Big Ear, designed and built by John D. Kraus in Delaware, Ohio, focused radio waves from a tiltable flat reflector and a fixed paraboloid onto two feed horns. It did not track; it stared at a fixed declination while the sky drifted past. Each beam was about 8 arcminutes wide, so a point source took about 72 seconds to cross it, and because the horns were offset a genuine source appeared in one and then, roughly three minutes later, in the other.

The receiver was a 50-channel filter bank, each channel 10 kHz wide, centred on the hydrogen line. Every 12 seconds the IBM 1130 computer printed one character per channel. Ehman's 2010 report describes the encoding: the computer subtracted a running baseline, divided the result by the standard deviation of the noise over the previous 60 integration periods, and truncated the ratio to an integer. Zero printed as a blank, 1 through 9 as digits, and 10 through 35 as the letters A through Z. Each character is therefore a signal-to-noise ratio in units of the background standard deviation. The string 6EQUJ5 decodes as 6, 14, 26, 30, 19, 5. The peak character, U, means the signal was between 30 and 31 times the noise; Ehman's best estimate is 30.5 ± 0.5 sigma.

The detection occurred at about 22:16 EST on 15 August 1977 (11:16 p.m. local daylight time), with nobody at the telescope. Ehman found it on the printout a few days later, probably on 19 August.

Frequency is where the record becomes complicated. Kraus quoted 1420.3556 MHz in a 1994 letter to Carl Sagan. Ehman, finding that the first local oscillator had been set 0.1 MHz above its nominal value, corrected this in 1997 to 1420.4556 ± 0.005 MHz. In August 2025, Abel Méndez and colleagues at the University of Puerto Rico at Arecibo reprocessed the original printouts, found that the filter bank's channels had been numbered in reverse frequency order, and moved the value to 1420.726 ± 0.005 MHz. All three lie near the hydrogen rest frequency of 1420.4058 MHz; the 2025 value implies a blueshift of about 84 km/s relative to the Sun, or 74 km/s in the Local Standard of Rest.

Intensity is similarly contested. Ehman's report gives two flux densities that were never reconciled: 54 Jy by his own method and 212 Jy by chief observer Russ Childers. Méndez's 2025 recalibration, using calibrator records on the same printouts, yields a lower limit of 256 ± 63 Jy.

Position depends on which horn caught the signal, which the software did not record. Ehman's J2000 coordinates were RA 19h25m31s (positive horn) or 19h28m22s (negative horn), Dec -26°57', with errors of ± 10 seconds of time and ± 20 arcminutes. The 2025 re-analysis narrowed these to RA 19h25m02s or 19h27m55s, each ± 3 seconds.

Timeline

• 15 August 1977At about 22:16 EST the Big Ear records 6EQUJ5 in channel 2 while scanning a strip of sky in Sagittarius.
• Circa 19 August 1977: Ehman circles the sequence and writes "Wow!" in the margin.
• 1994Kraus, writing to Carl Sagan, gives the frequency as 1420.3556 MHz.
• 1995-1996Robert Gray and Kevin Marvel search the field with the Very Large Array without result.
• 1998The Big Ear is demolished to expand a golf course and build housing.
• 2016Antonio Paris proposes that comets 266P/Christensen and P/2008 Y2 (Gibbs) caused the signal.
• June 2017Paris reports 1420 MHz emission from 266P; the Ohio State observatory staff publish a rebuttal within days.
• November 2020Alberto Caballero identifies 2MASS 19281982-2640123 as the most Sun-like star in the error box.
• 21 May 2022Breakthrough Listen observes that star with the Green Bank Telescope and Allen Telescope Array; nothing is found.
• 16 August 2024Méndez, Ortiz Ceballos and Zuluaga post "Arecibo Wow! I," proposing a hydrogen-cloud flare triggered by a magnetar.
• 15 August 2025"Arecibo Wow! II" revises the frequency, flux and position using 75,000 pages of preserved Big Ear printouts.
• 9 June 2026Méndez, Dixon and Childers post a history of the Ohio SETI program and its archive of more than 40,000 narrowband transients.

The Evidence

The primary artefact is the printout, preserved by the volunteer North American AstroPhysical Observatory (NAAPO). Behind it lies a far larger body of paper: Marc Abel of Wright State University kept the 1977-1984 printouts and photographed more than 75,000 pages, which the Arecibo Wow! team later ran through optical character recognition.

Several properties are not in dispute. The signal was confined to a single 10 kHz channel and was at least 30 times the noise. Its rise and fall over six samples match a Gaussian fit to the beam pattern with a half-power width of about 37 to 39 seconds; Ehman is emphatic that this width is not the duration, which was at least the full 72 seconds of the transit. It did not appear in the other horn, so within the three-minute gap between beams the signal either switched on, switched off, or was never a fixed celestial source.

Ehman's report walks through the terrestrial candidates. Planets, asteroids and known spacecraft were not in the field. Satellite orbits were checked, and the 1420 MHz band is protected worldwide from transmission. Aircraft would have crossed the beam too fast to mimic a point source, and a ground transmitter picked up directly would not show the beam-shaped rise and fall. Ehman concludes that all terrestrial explanations were "either ruled out or seem improbable." The 2025 re-analysis adds that the only Ohio UHF television stations whose second harmonic could land in the band were off the air, that GOES solar records show no matching burst, and that across years of surrounding data no random channel excursion reached 7 sigma.

The Theories

What Doesn't Add Up

The single-horn detection is the central anomaly and no theory handles it comfortably. A fixed source should have appeared twice within about five minutes; it appeared once. An alien transmitter would have to switch off within three minutes of switching on, a natural flare would have to last more than 72 seconds but less than about 150, and a fault would have to counterfeit a beam profile.

The measured properties keep moving. Three published frequencies span 0.37 MHz, wider than 37 receiver channels, and three flux estimates span a factor of nearly five. Each revision rests on a documented instrumental correction, but all derive from the same sparse printout, and the 2025 authors note that a different calibrator reference would double their flux again. No single agreed value exists for either quantity.

The archival reversal is itself striking. For decades the Big Ear's data were believed lost; the 2025 paper states the team began with "no intention" of re-evaluating the signal until volunteers surfaced the photographed pages. The June 2026 history reports more than 40,000 narrowband transients logged over the program's life, most unexamined. Whether the Wow! signal is unique or merely the strongest of a population is a question the archive could answer and has not yet been asked.

Latest Developments

Nearly all recent work comes from the Arecibo Wow! project led by Méndez. The 2024 paper introduced the hydrogen-cloud hypothesis and, as Méndez told Science News, was born of surprise: he had expected the Arecibo scans to show nothing and instead found cloud signals shaped like fainter copies of the original. The SETI Institute discussed the paper in January 2025, calling it a plausible natural explanation that the event's transient nature makes hard to confirm. The project page states the paper is being refined for journal submission.

The August 2025 paper delivered the first re-analysis of the original data with modern methods. Beyond the revised frequency, flux and position, it found "no evidence that the Wow! Signal was induced by external or internal interference" and results that "align with an astrophysical origin." Méndez said the work "doesn't close the case" but "reopens it, but now with a much sharper map in hand." The eleven authors include David Kipping; the paper remains a preprint as of this writing.

The June 2026 paper by Méndez, Dixon and Childers, submitted to the proceedings of IAU Symposium 404, situates the signal within the whole Ohio SETI program and frames the preserved archive as an open resource. A citizen-science arm, Wow@Home, has also been launched.

No new radio search of the revised positions has been published. The 2025 authors note that earlier searches by Gray, Harp and Breakthrough Listen covered the new fields, though not necessarily at the same sensitivity. The 50th anniversary falls in August 2027.

Sources

1. The Big Ear Wow! Signal (30th Anniversary Report) — Jerry R. Ehman, bigear.org / NAAPO, 2010
2. Arecibo Wow! II: Revised Properties of the Wow! Signal from Archival Ohio SETI Data — Méndez et al., arXiv, 2025
3. Arecibo Wow! I: An Astrophysical Explanation for the Wow! Signal — Méndez, Ortiz Ceballos and Zuluaga, arXiv, 2024
4. The Ohio SETI Program -- The Last Decades — Méndez, Dixon and Childers, arXiv, 2026
5. Statement regarding the comet claim for the Wow! signal — Robert S. Dixon, NAAPO / Ohio State University SETI Program, 2017
6. Hydrogen Clouds from Comets 266P/Christensen and P/2008 Y2 (Gibbs) are Candidates for the Source of the 1977 Wow! Signal — Paris and Davies, Journal of the Washington Academy of Sciences (arXiv copy), 2016
7. Hydrogen Line Observations of Cometary Spectra at 1420 MHz — Antonio Paris, Journal of the Washington Academy of Sciences (arXiv copy), 2017
8. Comet claim for mysterious Wow signal sparks controversy — Astronomy Now, 2017
9. Comet Likely Didn't Cause Bizarre 'Wow!' Signal (But Aliens Might Have) — Space.com / Live Science, 2017
10. An approximation to determine the source of the WOW! Signal — Alberto Caballero, arXiv, 2020
11. Breakthrough Listen Search for the WOW! Signal — Breakthrough Listen, UC Berkeley SETI Research Center, 2022
12. Sun-like star ruled out as source of mysterious WOW! signal — Astronomy magazine, 2022
13. Could the "Wow" signal have originated from a stochastic repeating beacon? — Kipping and Gray, arXiv / MNRAS, 2022
14. The historic 'Wow!' signal may finally have a source. Sorry, it's not aliens — Science News, 2024
15. The "Wow!" Signal Gets An Update - It Was Even Stronger Than We Thought — Universe Today, 2025

The Hinterkaifeck Murders

Six people killed with a farm mattock; someone then fed the cattle for days before the bodies were found

The Hook

In the last days of March 1922, the farmer Andreas Gruber told people in Schrobenhausen that something was wrong at home. He had heard footsteps in the loft all night, gone up with a lamp and found nothing. In the fresh snow behind the house he had seen tracks leading to the machine shed, and none leading away. A Munich newspaper nobody in the district subscribed to had turned up on the property. A house key was missing. He refused offers of help. Within days he, his wife, his daughter, his two grandchildren and a maid who had arrived that afternoon were dead, and whoever killed them appears to have stayed on the farm, milking the cows and eating the bread, while the bodies lay under hay in the barn.

The Facts

Hinterkaifeck was a smallholding of roughly 17 hectares about 900 metres from the hamlet of Gröbern near Schrobenhausen, some 70 kilometres north of Munich, bordered by woodland and ten to fifteen minutes on foot from the nearest neighbour. The farm belonged to Viktoria Gabriel, née Gruber; her parents Andreas and Cäzilia Gruber lived there in retirement.

The six victims were Andreas Gruber (63), Cäzilia Gruber (72), their widowed daughter Viktoria Gabriel (35), Viktoria's children Cäzilia Gabriel (7) and Josef Gruber (born 7 September 1919, aged 2), and the new maid Maria Baumgartner (44 or 45, depending on the source). Karl Gabriel had been conscripted in 1914 and was recorded killed in northern France on 12 December 1914.

The family's history was public. In 1915 the Neuburg court convicted Andreas Gruber and Viktoria of incest for a relationship dated to 1907–1910; he received one year, she one month. In September 1919 the neighbour Lorenz Schlittenbauer, a widower who had been Viktoria's lover, denounced father and daughter again after doubting his own paternity of the newborn Josef. He withdrew the accusation, then reasserted it under oath; Gruber was acquitted on 31 December 1919 because Schlittenbauer's statements were deemed unreliable. Schlittenbauer nonetheless acknowledged paternity.

On Friday 31 March 1922, at about half past four in the afternoon, Maria Baumgartner arrived to take up her post, accompanied by her sister Franziska Schäfer, the last outsider to see the family alive. From the victims' clothing the court commission placed the killings between 8 and 10 p.m. Viktoria, her mother, her father and the seven-year-old Cäzilia were struck down one after another at the door between stable and barn, their bodies stacked and covered with half a metre of hay and a wooden door. Inside the house, Maria Baumgartner was killed in the maid's chamber and covered with a blanket; Josef was killed in his cot in the bedroom, the blow passing through the pram hood.

The bodies were found on the evening of Tuesday 4 April by Schlittenbauer, Jakob Sigl and Michael Pöll, after the mechanic Albert Hofner had spent hours on the farm repairing an engine without meeting anyone. Munich police under Kriminaloberinspektor Georg Reingruber received the report at 6:15 p.m., reached the village at 1:30 a.m. and began at dawn on 5 April. By then onlookers had walked through the scene. No fingerprints were taken. The court physician Johann Baptist Aumüller performed the autopsies on 6 and 7 April; the heads were removed for examination.

The files were formally closed in 1955; the last interviews took place in 1986. More than a hundred people were investigated. No one was ever charged.

Timeline

• 12 December 1914Karl Gabriel, Viktoria's husband, recorded killed in northern France.
• 7 September 1919Josef Gruber born; Viktoria names Lorenz Schlittenbauer as father.
• 10 September – 31 December 1919: Schlittenbauer's incest complaint leads to Gruber's arrest, a withdrawal, a sworn reassertion and acquittal.
• November 1920 – September 1921: Maid Kreszenz Rieger works at Hinterkaifeck, then leaves.
• 30 March 1922Andreas Gruber finds tracks in the snow leading to the machine shed and none leading away; he tells acquaintances of footsteps in the loft.
• 31 March 1922Maria Baumgartner arrives at about 4:30 p.m.; all six inhabitants killed between about 8 and 10 p.m.
• 1 April 1922Coffee sellers knock and get no answer; a passer-by notes the chimney smoking; Cäzilia Gabriel is absent from school.
• 4 April 1922Mechanic Albert Hofner works on the farm for hours; Schlittenbauer, Sigl and Pöll break in and find the bodies; Munich police notified at 6:15 p.m.
• 5–8 April 1922: Scene examined, autopsies performed, heads removed; funeral and 100,000-mark reward on 8 April.
• February 1923Farm demolished; a blood-stained mattock and a penknife found in hiding places.
• 25 February 1944Augsburg justice building destroyed in an RAF raid; case files and the mattock lost.
• 1955Files formally closed; last interviews in 1986.
• 2007Fifteen trainee detectives at the Fürstenfeldbruck police college produce a cold-case report.
• 31 March 2022Centenary tour and memorial service in Waidhofen.
• Spring 2025New memorial of stone steles inaugurated near the site, replacing the old wayside shrine.
• 17 March – 10 April 2026: Bavarian Main State Archive exhibits the surviving Munich police file.

The Evidence

Only five crime-scene photographs were taken: an exterior, two of the bodies in the barn (the second after Schlittenbauer had moved them while searching for Josef), one of the maid and one of the pram. No fingerprints were lifted, although the technique was standard by 1922.

The court commission's inspection record is the most reliable description of the scene. All exits were locked from inside; the discoverers entered through the machine shed, whose door was held only by a wedged piece of wood. In the loft above it investigators found hollows in the straw where a person could have lain, a knotted rope leading down into the machine room, and roof tiles displaced at two points giving a view over the whole yard. A hearing test by the judge, Wiesner, established that shouting in the stable could not be heard in the maid's chamber; the recorded assumption that the family was lured out by restless cattle therefore rests on nothing audible.

A pickaxe found in the feed trough was initially taken for the weapon, but the pathologist ruled it out on 7 April because of a small round wound on Viktoria's head. The true weapon, a Reuthaue or mattock that Andreas Gruber had made himself, was found in February 1923 during demolition, hidden in a false floor above the stairs. A protruding screw on the handle matched the wound. The State Archives catalogue records that the tool was destroyed in Augsburg on 25 February 1944.

The autopsies recorded crushed skulls on all victims. The State Archives catalogue notes seven blows to Cäzilia Gruber's head and strangulation marks on her neck; German Wikipedia attributes the strangulation marks to Viktoria and calls the sources unclear. All sources agree that the seven-year-old survived for at least two hours and tore out tufts of her own hair. The heads were removed and, according to the State Archives, handed to the Neuburg prosecutor and considered missing ever since; German Wikipedia and web.de state they were destroyed in the 1944 Augsburg bombing; English Wikipedia says they were sent to Munich. The bodies were buried without them.

Money remained in the house: German Wikipedia and web.de cite about 1,800 gold marks; the archives speak of a larger quantity of gold and silver coin, an empty purse and papers scattered on the bed. The cattle had been fed, watered and milked; the bread was gone and meat freshly cut. Hofner, the mechanic, was not interviewed until 1925, and the maid's sister not until 1930.

The Theories

What Doesn't Add Up

The hearing test undercuts the one recorded explanation for why four people went, one after another, to the barn door, two of them in night clothes.

The loft hollows and displaced tiles are read as a killer's hiding place, yet Andreas Gruber searched the building with a lamp and reported nothing. German Wikipedia records the alternative reading that the hollows were the meeting place of father and daughter, which would explain his silence but not the tracks in the snow.

Hofner's account of 4 April implies that someone was still on the farm that day. He heard the dog barking inside the locked house on arrival; when he left, the dog was tied up outside and the barn door stood open. That evening the dog was found injured beside the bodies and the barn door was closed again.

A passing craftsman, Michael Plöckl, reported that on the night after the killings the oven was lit, a stranger blinded him with a lamp and the smoke smelled foul. Nobody examined the oven.

Sources give the date as the evening of 31 March or the night of 31 March to 1 April; the court commission's window ends at 10 p.m. English Wikipedia says the killer lived with the corpses for three days; the archives count four days to discovery.

The pickaxe in the trough and a blood-stained iron band reportedly found during demolition were never mentioned again once the mattock was identified. If either was a second weapon, more than one attacker is implied.

Latest Developments

Nothing has changed in the status of the case, and given the loss of the mattock, the files and the skulls, nothing realistically can. What has changed is the memorial and the archival access.

For the centenary on 31 March 2022 the district of Neuburg-Schrobenhausen organised a guided walk from Waidhofen church to the site followed by a memorial service.

The concrete wayside shrine that had marked the site since the 1920s, inscribed to the family who fell to a "godless murderer's hand" on 31 March 1922, disappeared around 2022; a January 2025 report from Pfaffenhofen describes it as standing "until recently". The Waidhofen horticultural society then commissioned the Kühbach stonemason Kress to create a group of stone steles a few metres from the old spot, beside the weather spruce. Skeptix reported the new memorial as just inaugurated in late April 2025, and the municipal newsletter for September 2025 records the society tending the victims' grave, the "vanished" shrine and the new memorial. Guided walks to the site continue several times a year.

A University of Augsburg doctoral thesis on how the Hinterkaifeck story has been transmitted was defended in September 2024 and published in 2025. Most recently, from 17 March to 10 April 2026, the Bavarian Main State Archive in Munich showed the surviving Munich police file (Polizeidirektion München 8091b), the crime-scene photographs and the 1922 autopsy reports in a small exhibition prepared by the archives' training college.

Sources

1. Der sechsfache Mord in Hinterkaifeck (exhibition catalogue, PDF) — Generaldirektion der Staatlichen Archive Bayerns / Bayerisches Hauptstaatsarchiv, 2026
2. Hinterkaifeck murders — Wikipedia (English), 2026
3. Mordfall Hinterkaifeck — Wikipedia (German), 2026
4. Der Mythos Hinterkaifeck — hinterkaifeck.net, 2023
5. In aller Kürze... — hinterkaifeck.net, 2023
6. Mythos Hinterkaifeck – Auf den Spuren eines Verbrechens — Bayerisches Armeemuseum, 2016
7. Auf den Spuren des bestialischen Sechsfach-Mordes von Hinterkaifeck — Pfaffenhofen Today, 2022
8. Der Sechsfach-Mord von Hinterkaifeck: Fackel-Tour zum Ort des Grauens — Pfaffenhofen Today, 2025
9. Deutschlands Mordrätsel Hinterkaifeck: Früher Tatort eines grausamen Gemetzels – jetzt ein Touri-Treff — t-online, 2022
10. Der Fall Tannöd: Wer beging den sechsfachen Mord in Hinterkaifeck? — WEB.DE News, 2022
11. "Der Mythos Hinterkaifeck" bei Skeptics in the Pub Augsburg — Skeptix, 2025
12. Mitteilungsblatt Waidhofen Nr. 9/2025 — Verwaltungsgemeinschaft Schrobenhausen, 2025
13. Mordfall "Hinterkaifeck" und seine Andachtsstätte — Deutschland mal anders, 2020

Amelia Earhart

The Itasca's radio log records her last words at 8:43 a.m.; the largest air-sea search of its era found nothing

The Hook

The last thing anyone heard from Amelia Earhart was a navigational line, not a distress call. At about 8:43 a.m. on 2 July 1937, the radio operator aboard the Coast Guard cutter Itasca, anchored off Howland Island, wrote down a fragment in his log: "we are on the line 157 337 wl rept msg we wl rept." The signal came in at strength 5 — clear, strong, close. Then nothing.

Everything about the transmission suggests she was near. The Itasca had been listening to her voice grow steadily louder for six hours. She was not lost in a storm. Visibility was good. She simply could not find an island two miles long, and the ship sitting beside it could not talk back to her.

The United States government then conducted the largest search it had ever mounted for a missing aircraft. It found nothing. Eighty-nine years later, that is still true.

The Facts

Amelia Earhart, 39, was the most famous aviator in the world and, by her own account, not the best pilot in it. Her navigator, Fred Noonan, was a former Pan American Airways master navigator with extensive Pacific experience.

They left Miami on 1 June 1937 in a Lockheed Electra 10-E, an all-metal twin-engine aircraft bought for her by the Purdue Research Foundation through the Amelia Earhart Fund for Aeronautical Research. Earhart called it her "Flying Laboratory" and intended to return it to Purdue after the flight. Her radio call sign was KHAQQ.

By 2 July they had flown more than 22,000 miles. Three ocean legs remained: Lae in New Guinea to Howland Island, Howland to Hawaii, Hawaii to California. Howland is roughly two miles long and one mile wide, a flat coral speck 2,556 miles from Lae. The US government had built an airstrip there and stationed the Itasca offshore with fuel.

They departed Lae at 10:00 a.m. local time on 2 July. Earhart expected to fly a minimum of eighteen hours and arrive around sunrise.

From 2:45 a.m. to 8:45 a.m. local time on 2 July, across the date line, the Itasca's operators received her intermittent voice messages. Signal strength rose progressively — she was getting closer. Two-way contact was never established. The Smithsonian's National Air and Space Museum attributes this to a combination of confused communications plans and the fact that neither Earhart nor Noonan could use Morse code, which the Itasca's crew did not know. It appears she may never have heard the ship at all.

Her strongest transmission came at 7:58, at strength 5-plus: she was circling, searching, and running low on fuel. The last, at 8:43, was strength 5.

The Navy and Coast Guard searched from 2 to 18 July. The battleship USS Colorado launched floatplanes from 7 July; the aircraft carrier USS Lexington took over the search on 12 July, concentrating north and west of Howland, and efforts were suspended on 18 July. Contemporary and later accounts put the total area covered at roughly 250,000 square miles, though that figure is attached in some sources to the Lexington phase alone. Nothing was found.

Timeline

• 1 June 1937Earhart and Noonan leave Miami on the second, eastbound attempt at a round-the-world flight.
• 2 July 1937, 10:00 a.m. local: The Electra departs Lae, New Guinea, for Howland Island.
• 2 July 1937, 2:45 a.m. local (Howland): The Itasca begins receiving intermittent voice transmissions from KHAQQ.
• 2 July 1937, 7:58 a.m.: Strongest signal received, strength 5-plus. Earhart reports circling and low fuel.
• 2 July 1937, 8:43 a.m.: Last recorded transmission — "we are on the line 157 337."
• 2-18 July 1937: US Navy and Coast Guard conduct the largest search for a lost aircraft to that date. No trace recovered.
• 1940Human bones, a sextant box and other items are found on Gardner Island, now Nikumaroro, in the Phoenix Islands.
• 1941Physician D. W. Hoodless examines the bones in Fiji and concludes they belonged to a male. The bones are subsequently lost.
• 1988 onward: The International Group for Historic Aircraft Recovery (TIGHAR) begins fieldwork on Nikumaroro, advancing what becomes known as the Nikumaroro hypothesis.
• 2002-2009Nauticos and, later, the Waitt Institute conduct deep-water sonar searches west of Howland based on the crash-and-sink model. Nothing is found.
• March 2018Richard Jantz publishes a re-analysis of the 1940 bone measurements in Forensic Anthropology.
• 27 January 2024Deep Sea Vision announces a plane-shaped sonar anomaly some 16,000 feet down, roughly 100 miles from Howland.
• 1 November 2024Deep Sea Vision returns to the site. High-resolution imaging shows a natural rock formation.
• 2025Purdue Research Foundation and the Archaeological Legacy Institute announce the Taraia Object Expedition, then postpone it pending approval from the government of Kiribati.
• 24 July 2026Purdue announces Kiribati cabinet approval; the expedition is confirmed to depart in early October 2026.

The Evidence

The radio log. The single most important document in the case is a Coast Guard radio log kept by operator Leo G. Bellarts, now held by the US National Archives as part of the Bellarts Papers. It records the crew reporting radio trouble, flying low at 1,000 feet, and being 200 miles from Howland, and ends with the 8:43 fragment. The "157 337" numbers are a reciprocal bearing — a celestial line of position running roughly northwest to southeast, the kind a navigator establishes from a sun sight when he knows his distance along a line but not his position on it. Howland lay somewhere on that line. So did open ocean, in both directions, for hundreds of miles. Following it southeast leads eventually toward the Phoenix Islands.

The Nikumaroro bones. In 1940, a British colonial officer recovered a partial skeleton on Gardner Island, an uninhabited coral atoll about 350 nautical miles southeast of Howland, along with a sextant box. The bones were sent to Fiji, where in 1941 Dr D. W. Hoodless measured them and concluded they came from a male. The bones were later lost; only Hoodless's measurements survive.

The artefacts. TIGHAR expeditions to Nikumaroro since 1988 have recovered objects including part of a woman's shoe, a compact case, a jar consistent with 1930s freckle cream and a medicine vial. None bears an identifying mark tying it to Earhart. The Bevington Object is an anomaly visible on the reef in a photograph taken about three months after the disappearance, which TIGHAR and now the Archaeological Legacy Institute argue resembles a landing gear assembly.

The Taraia Object. A visual anomaly in the Nikumaroro lagoon, first noticed in satellite imagery in 2020. The Archaeological Legacy Institute states that contemporaneous photographic evidence appears to place an object in the same location as early as 1938. No one has yet examined it in person.

The 2024 sonar target. In January 2024 the exploration company Deep Sea Vision announced sonar imagery of an aircraft-shaped object more than 16,000 feet down, roughly 100 miles from Howland. The team returned on 1 November 2024 and imaged it at close range. It is a rock formation.

The Theories

What Doesn't Add Up

The communications failure is the most inexplicable part of the case, and it was preventable. Earhart had removed the long trailing wire antenna needed for effective transmission on 500 kHz. She and Noonan could not read Morse. The frequency and schedule arrangements with the Itasca were never properly agreed. A pilot attempting the hardest over-water navigation leg of a record flight arrived at the critical moment with a radio setup she could transmit on but apparently not receive on.

The 157/337 line is a genuine navigational puzzle. A line of position tells you where you are along a line, not where you are on it. If Noonan had a good sun line, the correct procedure was to fly it in one direction and then the other. Which way he chose determines everything, and no one knows.

The fuel arithmetic has never been settled. Estimates of the Electra's remaining endurance at 8:43 a.m. range from minutes to more than an hour, which is the difference between the ocean near Howland and an atoll 350 miles away.

Finally, there is a persistent problem with the post-loss radio signals. In the days after the disappearance, amateur and official listeners across the Pacific reported hearing signals attributed to KHAQQ. Some were certainly hoaxes or misidentifications. A subset was taken seriously at the time and used to direct search assets. If any were genuine, the aircraft was on land with at least one engine able to run a generator — which the Colorado's floatplanes, flying over the Phoenix Islands on 9 July, did not detect.

Latest Developments

The most concrete recent development was negative. The Deep Sea Vision anomaly, announced in January 2024 as a possible Electra and covered worldwide, was re-imaged by the company's own autonomous vehicle on 1 November 2024 and identified as a natural rock formation. The company said the search continues.

Attention has since moved to Nikumaroro. In 2025 the Archaeological Legacy Institute, the Purdue Research Foundation and Purdue University announced the Taraia Object Expedition, intending to reach the island in November 2025. It was postponed while approvals were sought from the government of Kiribati, which administers Nikumaroro as part of the Phoenix Islands Protected Area.

On 24 July 2026 — Earhart's 129th birthday — Purdue announced that the Kiribati cabinet had approved the expedition. A sixteen-person team led by Richard Pettigrew, executive director of the Archaeological Legacy Institute, will assemble in Tarawa on 6 October 2026, depart by sea on 7 October, sail roughly 830 nautical miles to Nikumaroro and spend several days examining the Taraia Object directly. Three Purdue representatives are aboard, including the university's senior vice president and general counsel, Steve Schultz. If the object proves to be aircraft wreckage, the institute has said a further field season would follow, with possible recovery in 2027 or 2028.

As of September 2026 the expedition has not yet sailed. No physical object recovered anywhere has ever been conclusively identified as coming from Earhart's Electra.

Sources

1. Radio Log of the Last Communications of Amelia Earhart — US National Archives, 2024
2. Amelia Earhart: Missing for 80 Years But Not Forgotten — Smithsonian National Air and Space Museum, 2017
3. ALI, Purdue expedition to search for Amelia Earhart's plane will depart in October — Purdue University, 2026
4. The Taraia Object Expedition — Purdue Research Foundation / Archaeological Legacy Institute, 2026
5. Amelia Earhart and the Nikumaroro Bones: A 1941 Analysis versus Modern Quantitative Techniques — Forensic Anthropology (University of Florida Press), 2018
6. Forensic analysis of bones indicates that they are Amelia Earhart's — Phys.org / University of Tennessee, 2018
7. The Earhart Project — The International Group for Historic Aircraft Recovery, undated
8. Earhart Project Hypothesis — The International Group for Historic Aircraft Recovery, 2009
9. Explorers may have found Amelia Earhart's plane — CNN, 2024
10. Sonar imagery that looked like Amelia Earhart's plane was actually a natural formation of rocks — CNN wire / KRDO, 2024
11. Naval Aviation and the Search for Amelia Earhart — National Naval Aviation Museum, 2020
12. The Search for Amelia Earhart — Smithsonian Magazine, 2021
13. Top 3 Theories for Amelia Earhart's Disappearance — National Geographic, 2017
14. What Happened to Amelia Earhart? — HISTORY, updated 2025

Cicada 3301

Three cryptographic recruitment puzzles, fourteen posters on four continents, and a 58-page runic book that is still 96 percent unsolved

The Hook

The test for whether a message is really from Cicada 3301 is not stylistic. It is arithmetic. Every authentic communication the group has ever issued carries a signature from one OpenPGP key, fingerprint 6D85 4CD7 9333 22A6 01C3 286D 181F 01E5 7A35 090F, created on 3 January 2012. The key's user ID is "Cicada 3301 (845145127)". That number is not decorative: 845145127 is 3301 multiplied by 509 and by 503, all three prime, and 509 by 503 were the pixel dimensions of the image that started everything.

Cicada told solvers to check the signature, twice, in signed messages. The instruction has outlived the group. Since 4 April 2017 — the date of the last message that passes the test — nothing has. Dozens of puzzles have claimed the name. None verifies.

What is left is a 58-page book written in Anglo-Saxon runes, of which two pages have been read. The other fifty-six have resisted a community of thousands for twelve years.

The Facts

On or about 4 January 2012, an image appeared on 4chan: white text on black. "Hello. We are looking for highly intelligent individuals. To find them, we have devised a test. There is a message hidden in this image. Find it, and it will lead you on the road to finding us. We look forward to meeting the few that will make it all the way through. Good luck." It was signed 3301.

The date is genuinely disputed. Wikipedia and Fast Company give 4 January 2012; other walkthroughs give 5 January. The board is disputed too: The Register and Boxentriq say /x/, other accounts say /b/ or /sci/. The image was almost certainly reposted, and the original posting has never been established.

Solvers who ran the file through a JPEG steganography tool called OutGuess found a chain of stages: a Caesar cipher keyed to Claudius being Rome's fourth emperor, a decoy image of a duck, a book cipher keyed to Thomas Bulfinch's retelling of the Welsh Mabinogion, and eventually a countdown clock on a website named for that prime product.

At 17:00 UTC on 9 January 2012 the countdown expired and published fourteen sets of GPS coordinates under a cicada logo, with the instruction: "Find our symbol at the location nearest you." The coordinates resolve to roughly ten distinct localities across four continents — Warsaw; two adjacent points in Paris; three in Seattle; two in the Seoul area; Chino, California; the New Orleans area; Miami; Oahu; Erskineville in Sydney; and Fayetteville, Arkansas. At each, a paper poster bearing a cicada and a QR code. Scanning led to further steganographic images and further ciphers, one of them keyed to William Gibson's 1992 poem "Agrippa (A Book of the Dead)", a text famously distributed on a diskette programmed to erase itself after a single reading. That cipher produced a Tor hidden service address.

The 2012 round ended with a message on that hidden service stating that 3301 was disappointed to find solvers had collaborated, and that its winners had already been chosen.

A second puzzle followed a year later, in early January 2013, with eight poster locations. A third came in early January 2014, announced from the Twitter account @1231507051321. That round produced Liber Primus.

There was no 2015 puzzle, no 2016 puzzle, and no puzzle since.

Timeline

• 3 January 2012The OpenPGP key 7A35090F is created.
• 4 or 5 January 2012 (disputed): The first image is posted to 4chan.
• 9 January 2012, 17:00 UTC: Fourteen GPS coordinates are published; posters are recovered at ten localities on four continents.
• Early 2012The first puzzle closes; winners are admitted to a private forum.
• March 2013Marcus Wanner, one of the recruits, is told by another winner that the group has been "laid off". The project site is removed.
• 4 or 5 January 2013 (disputed): The second puzzle begins, with eight poster locations.
• Early January 2014: The third puzzle begins, announced via @1231507051321.
• May 20143301 releases the Liber Primus file dump — 58 pages, named 0.jpg to 57.jpg.
• 28 July 2015A signed message disavows a hacker group using the name "3301" in connection with a Planned Parenthood breach.
• 31 December 2015 / 5 January 2016: A signed message is published via the Twitter account: "The path lies empty; epiphany seeks the devoted. Liber Primus is the way... Beware false paths. Verify OpenPGP 7A35090F."
• 4 April 2017, 23:23 GMT: The last verified message is signed and posted to Pastebin: "Beware false paths. Always verify PGP signature from 7A35090F. 3301". It is not noticed until 29 April.
• August 2023Organisers of the CicadaSolvers community present "Cracking Cicada 3301" at DEF CON 31.
• April 2026Reporting on the fourteenth anniversary of the Fayetteville poster confirms the puzzle remains unsolved and the CicadaSolvers Discord has passed 10,000 members.

The Evidence

The key. Everything in this case reduces to one cryptographic fact. The 7A35090F key is a 4096-bit RSA key created the day before the first image appeared. Six messages have been signed with it. Nothing signed with it has appeared in more than nine years. Because the private key has never leaked, any future message bearing a valid signature would be strong evidence that whoever ran the puzzle is still alive and in possession of it.

The posters. Community documentation records that the website published fourteen coordinates, but the steganographic payload inside the image itself contained only twelve — the Fayetteville point and one Seoul point were missing from the file. Posters were nonetheless retrieved from both. Two poster variants existed. The Fayetteville and Erskineville posters are logged as found but no photograph of either survives; the 2012 round is poorly documented and parts of the record are lost. A widely circulated theory, popularised by a YouTube video, holds that no poster was more than about an hour from an airport, implying a small team travelling rather than a distributed international membership. Administrators of CicadaSolvers say years of analysis have produced no evidence for or against it.

Liber Primus. The book is written in a 29-rune Anglo-Saxon futhorc. Cicada supplied the key to it — a table called Gematria Primus assigning each rune a Latin value, a decimal value from 0 to 28, and a prime value from 2 to 109. The resulting language, which solvers call Runeglish, collapses U with V, C with K and S with Z, and lets single runes expand into digraphs, which is precisely why statistical attacks are so difficult.

The solved pages show the method. One is a plain transliteration. One uses Atbash against the reversed rune alphabet. One uses a Vigenère cipher keyed to the word DIVINITY, complicated by a rule that the key position does not advance at certain positions, breaking the repeating rhythm that normally betrays Vigenère. One uses a running shift derived from Euler's totient function on successive primes.

The counts are stable across sources. Liber Primus 1 is seventeen pages and is entirely solved. Liber Primus 2 is fifty-eight pages, of which two — the last two, 56.jpg and 57.jpg — were solved within days or weeks of release in 2014. Nothing has fallen since. Fifty-six pages remain.

The missing page. Page 56.jpg instructs the reader to find a deep-web page identified only by its SHA-512 hash, calling it "the duty of every pilgrim to seek out this page." It has never been found. Some solvers call it the fifty-ninth page.

What Cicada said it wanted. The stated aim, verbatim, was "highly intelligent individuals". Several 2012 winners told Rolling Stone that 3301 said it normally recruited by other means and had built the puzzles specifically to find cryptography and computer-security skill. Marcus Wanner, who was fifteen when he solved the 2012 puzzle, says the recruits were set to build CAKES — Cicada Anonymous Key Escrow System, a dead-man's switch to publish sensitive material automatically if a whistleblower were killed or imprisoned. The group dissolved before it was finished.

The Theories

What Doesn't Add Up

Nobody knows who made it. Fourteen years, three global puzzles, a physical operation spanning four continents, and no participant, contractor, printer or traveller has ever been identified with any evidence that survives scrutiny. Attributions have been made publicly — one man has been repeatedly named online since 2017 and denies authorship — but no claim about who ran the 2012 to 2014 puzzles has ever been tested against the only evidence that would settle it, which is possession of the private key.

The stopping point is odd. The 2016 message says "Liber Primus is the way. Its words are the map, their meaning is the road, and their numbers are the direction." That is an instruction to keep going. The 2017 message says only to check signatures. If the group intended the book to be read, it left no further help. If it did not, the 2016 message was a misdirection issued under its own verified key.

The record itself is deteriorating. Photographs of two of the fourteen posters do not exist. Several of the onion services are dead and unarchived. The community wiki has had to correct folklore that had already entered secondary sources — the compression noise on the 2016 image that circulated as a hidden QR code, the "fourteen cities" that were really ten localities, a Fast Company country list contradicted inside its own article.

And there is a structural clue nobody has been able to use. Solvers have observed that two identical runes appear consecutively in the unsolved text less often than chance predicts, implying the encryption of each rune depends on the ciphertext of the one before it. It is the same class of weakness that broke Enigma, where a letter could never encrypt to itself. Twelve years on, it remains an observation.

Latest Developments

There has been no cryptanalytic breakthrough. The solved-page count has not moved since 2014.

In August 2023, four CicadaSolvers organisers presented "Cracking Cicada 3301: The Future of Collaborative Puzzle-Solving" in the War Stories track at DEF CON 31, a talk about sustaining a volunteer effort across nine years of failure rather than about the cipher itself.

Activity through 2024 was journalistic and retrospective: new walkthrough instalments, a television episode in September 2024, essays on whether machine learning could help. None of it constituted progress.

In 2025 a lawsuit was filed in the Michigan Court of Claims by a man long publicly associated with the Cicada name against an academic and her university over a published paper. The dispute concerns a separate attribution about online political activity and has nothing to do with the cipher; it will not establish who authored the puzzles. It is noted here only because it is frequently misreported as a Cicada development.

The most recent substantive reporting located is an April 2026 feature marking the fourteenth anniversary of the Fayetteville poster. It confirms the position: no puzzle since 2014, no verified message since 4 April 2017, the CicadaSolvers Discord past 10,000 members, and fifty-six runic pages still unread.

Sources

1. Cicada 3301 — Wikipedia, accessed 2026
2. Cicada: Solving the Web's Deepest Mystery — Rolling Stone, 2015
3. Cicada 3301: The web's toughest and most creepy crypto-puzzle is BACK — The Register, 2014
4. Meet The Man Who Solved The Mysterious Cicada 3301 Puzzle — Fast Company, 2014
5. After over a year of silence, a mysterious online puzzle has returned with a cryptic tweet — Business Insider, 2016
6. Cicada 3301 First Puzzle Walkthrough — Boxentriq, 2026 revision
7. Cicada 3301 Liber Primus Guide — Boxentriq, 2026 revision
8. Liber Primus Cryptanalysis Briefing — CicadaSolvers, 2026 (community-maintained)
9. 2016 Message — Uncovering Cicada Wiki, undated (community-maintained)
10. PGP Signed Message April 2017 — Uncovering Cicada Wiki, undated (community-maintained)
11. Map of all locations of 3301's posters — Uncovering Cicada Wiki, undated (community-maintained)
12. Liber Primus Unsolved Pages — Uncovering Cicada Wiki, undated (community-maintained)
13. A flyer linked the UofA to one of the internet's deepest mysteries. 14 years later, it's still unsolved — The Arkansas Traveler, 2026
14. MSU professor's QAnon paper prompts lawsuit with Cicada 3301 puzzle leader — The State News, 2025

Frederick Valentich

A 20-year-old pilot told Melbourne "it is hovering and it's not an aircraft"; seventeen seconds of open microphone followed

The Hook

The last words Frederick Valentich transmitted were not "it's not an aircraft." He had said that twenty seconds earlier. His actual final transmission was his own call sign and the station he was calling: "Delta Sierra Juliet Melbourne." Then the official Department of Transport transcript records a notation rather than a word — seventeen seconds of open microphone — and the line went dead.

He was twenty years old, 150 hours into a flying career that was not going well, alone in a rented Cessna over one of the least forgiving stretches of water in the southern hemisphere, in perfect weather, twenty-five minutes before he was due at King Island. He had taken four life jackets for a flight he told two different people two different reasons for making.

Forty-eight years later the aircraft has not been found, and in 2025 the Victorian government listed it as a wreck it would like to locate.

The Facts

Frederick Paul Valentich, born 9 June 1958, held a private pilot licence issued in September 1977. On 11 May 1978 he obtained a Class Four instrument rating, which authorised night flight only in visual meteorological conditions. He was not qualified to fly in cloud. He had roughly 150 hours total.

He was studying part-time for a commercial licence. He had failed all five theory subjects twice and three of them a third time; departmental files record that he was running out of time. He had been warned once for straying into controlled airspace at Sydney and cited twice for deliberately flying into cloud, with prosecution under consideration when he disappeared. He had twice been rejected by the RAAF on educational grounds and was a member of the Air Training Corps.

On the afternoon of Saturday 21 October 1978 he filed a flight plan at the Moorabbin briefing office at 17:23 for a night VMC flight from Moorabbin to King Island and back, routing via Cape Otway, below 5,000 feet: 41 minutes to Cape Otway, a further 28 to King Island. He refuelled to capacity at 18:10 — a total endurance of 300 minutes against a 69-minute leg. He took four life jackets. He made no arrangement for King Island's aerodrome lighting to be switched on for his arrival, which a night landing there would have required.

He told the briefing officer and the operator's representative he was collecting friends at King Island. He told others he was going for crayfish. The official report records that no evidence was found that either was true.

The weather was, in the report's words, clear with a trace of stratocumulus at 5,000 to 7,000 feet, scattered cirrus at 30,000, excellent visibility and light winds. Sunset was 18:43; end of daylight at Cape Otway was 19:18. A Bureau of Meteorology letter on file described conditions as perfect for night flying and noted that an aircraft over King Island at 19:00 could clearly see the Cape Otway lighthouse.

He departed Moorabbin at 18:19 and reported Cape Otway at 19:00. Six minutes later he called Melbourne Flight Service and asked whether there was any traffic below 5,000 feet. The flight service officer, Steve Robey, replied that there was none.

The SAR Alert Phase was declared at 19:12 and the Distress Phase at 19:33 when the aircraft failed to arrive.

Timeline

• 17:23, 21 October 1978: Valentich files a night VMC flight plan at Moorabbin for King Island and return.
• 18:10: The Cessna is refuelled to capacity — 300 minutes of endurance for a 69-minute leg.
• 18:19: VH-DSJ departs Moorabbin.
• 18:43: Sunset.
• Approximately 18:47: Roy Manifold photographs the sunset over Bass Strait from Crayfish Bay near Cape Otway lighthouse.
• 19:00: Valentich reports Cape Otway.
• 19:06:14: He asks Melbourne Flight Service about traffic below 5,000 feet. At 19:06:26 Robey replies "no known traffic".
• 19:06-19:12: The encounter sequence. He reports four bright lights, an object orbiting above him, a green light and a shiny metallic surface, a rough-idling engine, then that the object is hovering and is not an aircraft.
• Approximately 19:12: Final transmission — "Delta Sierra Juliet Melbourne" — followed by seventeen seconds of open microphone. Alert Phase declared.
• 19:33: Distress Phase declared.
• 22 October 1978An oil slick is found about 18 miles north of King Island.
• 25 October 1978The air, sea and land search — an RAAF P-3 Orion, eight civilian aircraft and shipping, covering more than 1,000 square miles — is called off after four days without result.
• 6 April 1979The original air traffic control tape is recycled and returned to service.
• 1983An engine cowl flap washes ashore in Bass Strait and is matched to a range of Cessna 182 serial numbers.
• 27 April 1982The Aircraft Accident Investigation Summary Report is approved for publication.
• 2012-2016Researcher Keith Basterfield obtains the departmental files through the National Archives of Australia and a freedom-of-information request to the ATSB.
• 19 December 2023ABC Radio National broadcasts "Last Light", with interviews including Steve Robey.
• 2025Heritage Victoria completes a survey of submerged aircraft wrecks in Victorian waters and lists VH-DSJ as a target for future search.

The Evidence

The transcript. The Department of Transport transcript carries a header note that words in brackets are open to other interpretations — the department itself flagged that parts of the tape were ambiguous. Much of what circulates online as verbatim quotation is one of several possible readings.

What the official document does contain: "I cannot affirm it is four bright it seems to be like landing lights"; "it seems to me that he's playing some sort of game he's flying over me three times at a time at speeds I could not identify"; "it's not an aircraft it is" followed by two seconds of open microphone; "it seems like it's stationary what I'm doing right now is orbiting and the thing is just orbiting on top of me also it's got a green light and sort of metallic (like) it's all shiny (on) the outside"; "it's just vanished"; "the engine is is rough idling I've got it set at twenty three twenty four and the thing is (coughing)". Asked his intentions, he replied: "my intentions are ah to go to King Island ah Melbourne that strange aircraft is hovering on top of me again — it is hovering and it's not an aircraft."

The report then records "DELTA SIERRA JULIET MELBOURNE // 17 seconds open microphone //" and notes that there is no record of any further transmissions.

Two corrections to the popular account. First, the report's transcript does not describe that seventeen seconds as a sound at all — the "metallic" characterisation comes from Department of Transport personnel quoted in The Age on 23 October 1978 and from wire copy the same day. The Civil Aviation Safety Authority's magazine, working from the files, calls it an unidentified staccato noise that ended abruptly. Second, "twenty three twenty four" is an engine power setting, not a time or a coordinate.

The tape no longer exists. In 1985 a television production company sought the unedited recording under freedom of information. The Bureau of Air Safety Investigation replied that departmental practice was to recycle tapes and that this one had been returned to service — on 6 April 1979, per a subsequent letter. No authenticated recording of the exchange is held by the Australian government. Any audio circulating as "the original" is of unverified provenance.

The controller. Steve Robey has never accepted the mundane explanations. Interviewed in 1999: "I don't believe he was disorientated. I don't believe he faked it or committed suicide. He sounded under stress and sounded very genuine." In later remarks he described Valentich as "not to the point where he was panicking, but he was genuinely concerned."

The search. Four days, more than 1,000 square miles, an RAAF Orion and eight civilian aircraft. Nothing. An oil slick found on 22 October about 18 miles north of King Island was never linked to the aircraft; testing reported by the Sydney Morning Herald concluded it came from a marine diesel engine. A month later the pilot of another Cessna reported the outline of what he took to be a submerged aircraft but could not relocate it on a second pass. A 1978 departmental report surfaced in 2012 states that search aircraft sighted and then lost a debris field; no follow-up document has been located.

The cowl flap. In 1983 an engine cowl flap washed ashore — on Flinders Island according to Wikipedia citing a National Archives document, the Tasmanian Aviation Historical Society and the Aviation Safety Network; on King Island according to Flight Safety Australia in 2025. The two are about 250 kilometres apart at opposite ends of Bass Strait, and the discrepancy is unresolved. In July 1983 the Bureau of Air Safety Investigation asked the Royal Australian Navy Research Laboratory to assess whether the part could have drifted there from the loss area. Nothing conclusive was reported.

What BASI concluded, verbatim, was that "the part has been identified as having come from a Cessna 182 aircraft between a certain range of serial numbers." VH-DSJ falls within that range. That is a class match, not an identification: it establishes only that the aircraft is not excluded. The range has never been published, so the size of the candidate pool is unknown. This remains the only physical object ever tentatively associated with the case.

The Theories

What Doesn't Add Up

The official conclusion is a single sentence: "The reason for the disappearance of the aircraft has not been determined." The outcome was recorded as presumed fatal and Valentich was declared dead in 1982.

The unresolved details are not exotic. The exact time of the final transmission is given variously as 19:11:09, 19:12, 19:12:28 and 19:12:49 across sources that all claim to be quoting the same report. The report's own approval stamp reads 27 April 1982; secondary sources say May 1982, and one 2025 account says 1981. The cowl flap was found on Flinders Island or King Island depending on which source you read. The serial-number range that would tell you how many aircraft besides VH-DSJ fit the part has never been published.

And the primary evidence was destroyed as a matter of routine. The tape was wiped and put back into service five and a half months after the disappearance, while the investigation was still open. Everything anyone argues about — whether the voice sounded frightened or performed, whether the word was "stationary" or "chasing me", what the seventeen seconds actually contained — depends on a recording that no longer exists.

Latest Developments

The modern evidential base was assembled by the researcher Keith Basterfield. In 2012 he obtained the National Archives release of the main investigation file (series B1497, control symbol V116/783/1047), and in August 2012 the remaining twenty-three folios by freedom of information from the Australian Transport Safety Bureau — the release that established the tape had been recycled. In January 2016 he located a third file (series B638, M116/783/1047), which contains the largest known collection of contemporary press clippings and the Brunt meteor letter. Its inside cover refers to a fourth departmental file that does not appear in the National Archives catalogue and has never been found.

In December 2021 Snopes published a full documented transcription of the official report, now the most accessible reliable text of the transcript. George Simpson's book Nothing on Radar: The Valentich Mystery appeared in 2022.

On 19 December 2023, ABC Radio National's The History Listen broadcast "Last Light", presented by Kirsti Melville and produced with Patrick Stokes, with interviews including Steve Robey and Rhonda Rushton. In February 2025 Flight Safety Australia, the Civil Aviation Safety Authority's magazine, published a detailed analysis working from the departmental files and reframing the case through modern human-factors concepts: expectation bias, spatial disorientation and pilot distraction.

The most concrete institutional development came in 2025. Heritage Victoria completed a survey of all known aircraft wrecks in Victorian waters — 104 sites from 1918 onward — and formally listed VH-DSJ in the Australasian Underwater Cultural Heritage Database. The agency states that the disappearance "is arguably the best-known aircraft loss at sea in Australia" and that finding the remains would allow its maritime archaeologists to answer the question of what happened. The data is being folded into a programme to identify the most significant aircraft wrecks still to be found.

No search has been announced or funded. As of September 2026 there has been no reopening by the ATSB, no new archival release, and no confirmed wreckage.

Sources

1. Disappearance of Frederick Valentich — Wikipedia, accessed 2026
2. Leaving this world — Flight Safety Australia (Civil Aviation Safety Authority), 2025
3. The Valentich Disappearance: Another UFO Cold Case Solved — Skeptical Inquirer, 2013
4. Frederick Valentich's 'UFO' Sighting and Disappearance — Snopes, 2021
5. Mysteries of Aviation: Frederick Valentich and VH-DSJ — Tasmanian Aviation Historical Society, 2020
6. Valentich: the rest of the file is available — Keith Basterfield, 2012
7. Another Australian government Valentich file emerges — Keith Basterfield, 2016
8. Last Light: the Valentich disappearance — ABC Radio National, 2023
9. Nightlife History: The Valentich Mystery — ABC Radio Sydney, 2024
10. Submerged aircraft wrecks program — Heritage Victoria, 2026
11. Accident Cessna 182L Skylane VH-DSJ, 21 October 1978 — Aviation Safety Network, undated (user-contributed)
12. Transcript of Frederick Valentich's final communications with Melbourne air traffic control — UFO Research (NSW), 2015

The Sodder Children

Five children left no remains in a fire, and the man who had threatened their father sat on the coroner's jury

The Hook

The state issued five death certificates on 30 December 1945. Cause of death: "fire or suffocation." Nothing had been found to certify.

The search of the ruins on Christmas morning produced no bodies. Firefighters told the parents the blaze had burned hot enough to cremate five children completely, in a wood-frame house that had come down in under an hour. Four years later, when the family paid a Washington pathologist to dig the site properly, the only human bone recovered was four lumbar vertebrae from a single person — and the Smithsonian Institution's examiners reported that those vertebrae showed no sign of ever having been exposed to fire.

George and Jennie Sodder spent the rest of their lives on the assumption that their children had walked out of that house alive. They put up a billboard on Route 16 with the five faces on it and left it there for more than thirty years. It came down in 1989, when Jennie died.

The Facts

George Sodder was born Giorgio Soddu in Tula, Sardinia, in 1895, and arrived at Ellis Island in 1908 at the age of thirteen. He worked the Pennsylvania railroads, moved to West Virginia, and built a coal-hauling business. He married Jennie Cipriani, a shopkeeper's daughter who had come from Italy as a small child, and they had ten children between 1923 and 1943. They lived in a two-storey house outside Fayetteville, a coal town in the Appalachians with a substantial Italian immigrant population.

On the night of 24 December 1945, nine of the ten children were at home. The second-eldest son, Joe, was away in the Army.

At about 12:30 a.m. on Christmas morning, the telephone rang. Jennie answered. A woman's voice asked for a name she did not recognise; there was laughter and the sound of glasses in the background. Jennie said it was the wrong number. Going back upstairs she noticed the downstairs lights were still on, the curtains open and the front door unlocked — chores the older children had been told to do. She locked up and went back to bed. Shortly afterwards she heard a hard thud on the roof and a rolling sound. About half an hour later she woke to smoke.

George, Jennie and four children got out: John (23), Marion (17), George Jr. (16) and two-year-old Sylvia, whose crib was in her parents' room. Maurice (14), Martha (12), Louis, Jennie Irene (8) and Betty (5) did not. Louis's age is given as nine in some accounts and ten in others; the discrepancy has never been resolved in the published record.

What happened next is the core of the case. George broke a window to re-enter and tore open his forearm; the staircase was already burning. He went for the ladder he kept propped against the house and it was not there. He tried to drive one of his two coal trucks up to the house to climb on it and reach the upper windows; neither engine would start, though both had run the day before. He went to a rain barrel and found it frozen solid.

Marion ran to a neighbour's to telephone the fire department and could not raise an operator. A second neighbour who saw the fire tried from a tavern, also without success, and finally drove into town to find Fire Chief F. J. Morris. Fayetteville's alarm system was a telephone tree: one volunteer rang the next. The station was about two and a half miles away. A truck reached the house at roughly 8 a.m.

The fire was ruled accidental — faulty wiring in George's office — by a state police inspector, and a coroner's jury sitting the day after the fire accepted that finding.

Timeline

• 24 December 1945, late evening: the five younger children stay up with their Christmas presents.
• Approximately 00:30, 25 December: the wrong-number telephone call. Jennie finds the lights on, the curtains open and the door unlocked.
• Shortly after 00:30: a thud on the roof followed by a rolling noise.
• Approximately 01:00: Jennie wakes to smoke. George's office is burning. Six people escape; five children do not.
• Approximately 08:00: the fire department arrives. The house is gone.
• 25 December: a search of the ruins produces no identifiable remains. Morris tells the family the fire cremated the bodies.
• 26 December: a coroner's jury rules the fire accidental, caused by faulty wiring.
• 27 December: the fire marshal's investigation begins.
• Approximately 29 December: the basement is bulldozed and covered with about five feet of dirt. Accounts differ on who did it — George himself, or Jennie's brother.
• 30 December 1945five death certificates are issued, cause "fire or suffocation."
• 1947the Sodders write to the FBI. J. Edgar Hoover replies that the matter appears local and outside the Bureau's jurisdiction; agents offer to assist with local permission, which Fayetteville police and fire officials decline.
• 1947-1949private investigator C. C. Tinsley works the case.
• August 1949pathologist Oscar B. Hunter excavates the site. Four lumbar vertebrae are recovered and sent to the Smithsonian Institution.
• October 1949the Smithsonian report is returned. Two hearings follow at the State Capitol in Charleston.
• 1949-1950Governor Okey L. Patteson and State Police Superintendent W. E. Burchett tell the Sodders the search is hopeless. The case is closed.
• 1952the FBI looks into a lead placing the children with relatives in Cortez, Florida. Nothing is found.
• 1952-1953the billboard goes up on Route 16. Sources differ on the year; contemporary newspaper coverage of its erection dates from September 1953.
• Late 1960s: an envelope postmarked in Kentucky arrives containing a photograph of a young man and the note "Louis Sodder. I love brother Frankie. Ilil Boys. A90132 or 35."
• 1969George Sodder dies.
• 1989Jennie Sodder dies. The billboard comes down.
• 2021Sylvia, the youngest child and last survivor of the fire, dies at 79.
• 7 September 2026the case is re-examined publicly with the participation of a family descendant, with a fresh public appeal for information.

The Evidence

The absence of remains. This is the whole case. A wood house, a fire of somewhere between forty minutes and several hours depending on whose account is used, and five children who left nothing recoverable. Jennie Sodder tested the claim herself, burning chicken bones, beef joints and pork bones, and got charred bone every time. In 1947 the family obtained a statement from an Ohio crematorium: bodies held two hours at 2,000 °F leave bones intact and recognisable. Household appliances were recovered from the basement in identifiable condition.

The 1949 bones. Hunter's excavation produced damaged coins, a partly burned dictionary and four lumbar vertebrae. The Smithsonian's assessment: the vertebrae belonged to one individual aged roughly 16 or 17, with an upper limit near 22, which is older skeletal maturation than expected for a fourteen-year-old — possible for Maurice, but not probable. The report added that the bones showed no evidence of fire exposure, that finding only four vertebrae after a careful excavation was "very strange," and that one would expect the full skeletons of five children. Its conclusion was that the vertebrae most likely came in with the fill dirt George had used to bury the basement.

The ladder and the trucks. The ladder was later found down an embankment about seventy-five feet from the house. Neither coal truck would start. A neighbour reported seeing a man taking a block and tackle — the gear used to pull engines — from George's property. A local man, Lonnie Johnson, was identified and pleaded guilty; accounts differ on whether his plea covered the theft alone or the theft and the cutting of the telephone line. He denied any part in the fire.

The telephone line. Investigators established the line had been cut. A repairman told the family the line looked cut rather than burned.

The wiring. If faulty wiring caused the fire, the circuit should have failed. The family stated that the downstairs lights were burning brightly during the blaze. George also noted that the local power company had recently inspected the wiring and passed it.

The threat. Some months before the fire, a life-insurance salesman George turned down told him his house would go up in smoke and his children would be destroyed, as payment for the remarks he had been making about Mussolini. George's opposition to the Italian dictator was well known in Fayetteville. That same salesman sat on the coroner's jury that ruled the fire accidental the day after it happened. He had previously worked for George and, according to reporting by the Times West Virginian, was a co-signer on the family's home insurance, which he had raised from $1,500 to $1,750 without their knowledge. He was never investigated.

The liver. Tinsley heard that Morris, who had publicly reported finding nothing, had privately said he found a human organ in the ashes, put it in a dynamite box and buried it at the scene. The family dug up the box and took it to a funeral director, who identified the contents as beef liver, untouched by fire.

The sightings. A woman at a Charleston hotel said four children resembling the Sodders registered there about midnight with two men and two women "all of Italian extraction," that the men refused to let her speak to the children, and that the party left early the next morning. A woman running a tourist stop between Fayetteville and Charleston said she served the children breakfast the morning after the fire and noted a car with Florida plates. None was ever corroborated.

The Theories

What Doesn't Add Up

The coroner's jury sat on 26 December, one day after the fire, and included a man who had recently threatened to burn the house down. That alone should have voided the finding.

The official cause was faulty wiring, but the lights stayed on during the fire, and the wiring had recently been inspected and passed.

The fire chief reported no remains, then allegedly buried beef liver at the scene and told people it had been found there.

The FBI offered to help and the local authorities said no.

Sources contradict each other on several points and should be treated with care. The fire's duration is given as thirty to forty-five minutes in most accounts and as all night by Horn. The Christmas Day search is described in most accounts as finding nothing at all, but at least one account records that a few bones and internal material were found and dismissed. The billboard is dated to 1952 in some reporting and 1953 in the contemporary newspaper record. Louis Sodder's age is given as both nine and ten. None of these is decisive, but the case has been retold so often that repetition has hardened some details that were never solid.

The single hardest fact remains unexplained by any theory: the only human bone ever recovered from the site had not been in a fire.

Latest Developments

No official investigation into the disappearance of the Sodder children is currently open. The case was closed at the state level around 1950 and has never been formally reopened by West Virginia authorities.

Sylvia Sodder Paxton, the two-year-old carried out of the house and the last survivor of the fire, died in 2021 at the age of 79. She maintained to the end that her siblings had not died that night.

On 7 September 2026 the case was re-examined publicly in a widely distributed account produced with Miranda Weeks, a descendant of one of the older Sodder children, who supplied family photographs and documents not previously in general circulation and issued a fresh appeal for information. The material presented revisits the conduct of local officials and the absence of remains, and raises a possible connection to a former employer of George Sodder; that connection has not been tested by any investigative body and should be treated as an assertion rather than a finding.

Eighty-one years on, no remains of Maurice, Martha, Louis, Jennie Irene or Betty Sodder have ever been identified, and no person has ever been charged in connection with the fire.

Sources

1. What Happened to the Sodder Children, the Siblings Who Went Up in Smoke in a West Virginia House Fire? — Smithsonian Magazine, 2024
2. The mystery of the vanished Sodder children has never been solved, but the cold case has haunted this small town for decades — ABC News (Australia), 2023
3. UPDATE: The Sodder Children — Crime Junkie / audiochuck, 2026
4. Mystery of Missing Children Haunts W.Va. Town — NPR, 2005
5. LONG, LONG, LONG Sodder Post — Stacy Horn, 2005
6. The Sodder Children — Stacy Horn, 2005
7. Their fate: kidnapped, murdered, or are they still alive? — Times West Virginian, 2016
8. Missing Sodder Children in West Virginia — Legends of America, undated
9. Sodder children disappearance — Wikipedia, accessed 2026
10. Christmas house fire, West Virginia, missing children — The Independent, 2022
11. What Really Happened to the Sodder Children — West Virginia Regional Biographies / RootsWeb, undated
12. Internet sleuths awaken 61-year-old Christmas mystery — Montgomery Herald, 2006

The Isdal Woman

Eight false Belgian identities, every clothing label removed, and a coded notebook of her own movements

The Hook

The police had her fingerprints, her teeth, two suitcases of her belongings, a handwriting sample, a list of the hotels she had slept in and the dates she had slept in them. They had the coded notebook in which she had recorded her own itinerary, and within weeks they had broken the code. They had eight names she had used and the towns she claimed to have been born in.

Not one of the names existed. Not one of the addresses existed. With a single exception, the streets did not exist either.

Everything she owned had been stripped of anything that could identify where it came from. The labels were cut out of her clothes. The prescription information was scraped off her eczema cream. What she could not remove she did not carry.

Bergen police closed the case on 22 December 1970, three and a half weeks after she was found, and before the toxicology and autopsy reports had come back.

The Facts

On Sunday 29 November 1970 a man walking in Isdalen with his two young daughters found a woman's body among the scree on the eastern side of the valley, below Mount Ulriken outside Bergen. Isdalen — "Ice Valley" — has an old local nickname, Dødsdalen, "Death Valley," from medieval accounts of people dying there.

She was lying on her back with her head pointing downhill and her arms drawn up in the posture heat produces in a body. The front of her body and her clothing were severely burned; her face was not identifiable. Bergen police filed it as case 134/70.

Around her, investigators recorded an empty bottle of St. Hallvard liqueur, two plastic bottles smelling of petrol, a plastic passport holder, rubber boots, a woollen jumper, a scarf, nylon stockings, an umbrella, a purse, a watch, two earrings, a ring, a matchbox and the remains of burned paper. A fur hat with traces of petrol on it lay beneath the body. The jewellery had been removed and laid out beside her rather than left on her. Every label and maker's mark on her clothing and possessions had been removed.

The autopsy at the Gades Institute found soot in her airways: she was alive when the fire started. Cause of death was recorded as a combination of barbiturate poisoning and carbon monoxide. Analysis indicated she had taken a large quantity of the sleeping drug Fenemal (phenobarbital) — estimates in the case file run to fifty or more tablets in two or three doses — and further undissolved tablets were found beside the body. There was bruising on the right side of her neck. Her teeth showed extensive and distinctive work: multiple root fillings and gold crowns of a style Norwegian dentists did not use, which pointed abroad.

Three days later, police matched fingerprints from a pair of sunglasses in a left-luggage locker at Bergen railway station to the body. The locker held two suitcases.

Timeline

• 24 March 1970the earliest confirmed Norwegian hotel registration recovered by police.
• Spring and autumn 1970: she travels repeatedly through Norway — Oslo, Trondheim, Stavanger, Bergen — and abroad, including Paris, with a probable stay in Switzerland. Police reconstruct the itinerary from hotel registration cards.
• 22-29 October 1970: Paris.
• 18-23 November 1970: she stays at the Hotel Hordaheimen in Bergen, in room 407, under the name Elisabeth Leenhouwer.
• 23 November 1970she checks out, pays in cash, orders a taxi. This is the last confirmed sighting. What happened over the following six days is unknown.
• 29 November 1970the body is found in Isdalen.
• 2 December 1970two suitcases are recovered from a locker at Bergen railway station and linked to her by fingerprints.
• 22 December 1970Bergen police close the case. At a press conference the lead investigator says espionage can be ruled out "with certainty" and points to suicide, while acknowledging that laboratory results are still outstanding.
• 2 January and 7 January 1971: the toxicology and full autopsy reports are delivered, after the case has already been closed.
• 5 February 1971she is buried in an unmarked Catholic grave at Møllendal cemetery, Bergen, in a zinc coffin chosen specifically so that the remains could be recovered later. Sixteen Bergen police officers attend. No relative comes forward.
• 1 November 2016the Norwegian Institute of Public Health succeeds in extracting DNA from tissue samples preserved since the autopsy.
• January-May 2017: the Institute of Forensic Medicine at the University of Innsbruck, under Walther Parson, analyses the material and concludes she was of European origin.
• May 2017Kripos issues an international appeal through Interpol seeking a match.
• 8 January 2018NRK publishes the isotope results from her teeth.
• 26 March 2018the NRK and BBC World Service podcast Death in Ice Valley begins publication; it runs across two seasons into 2021.
• 12 June 2023the Neue Zürcher Zeitung publishes an investigation arguing she was killed and proposing a network connection.
• As of September 2026: no identification.

The Evidence

The suitcases. Clothing and shoes with the labels removed. Wigs. Cosmetics. A tube of eczema cream with the prescription details scraped away. Non-prescription glasses and the sunglasses that carried her fingerprints. Maps and timetables. Norwegian kroner plus Belgian, British and Swiss currency. Five one-hundred-Deutsche-Mark notes concealed in the lining. A sewing kit from the Hotel Régina in Geneva — the one item she had missed.

The notebook. It contained short strings of letters and numbers. Police decoded it against the hotel registration cards and found it was her own travel log: dates and city initials. "20 M 23 M O" — 20 to 23 March, Oslo. "O 22 O 28 P" — 22 to 28 October, Paris. The itinerary checked out against the paperwork. One line was never resolved: a string read as referencing 23 November, the day she checked out of her last hotel, followed by two sets of initials that may be the initials of two people.

The names. Eight identities appear on Norwegian hotel registration forms: Geneviève Lancier, Claudia Tielt (used twice with two different addresses), Claudia Nielsen, Alexia Zarna-Merchez, Vera Jarle, Fenella Lorck, Finella Lorck and Elisabeth Leenhouwer. Each claimed Belgian nationality. Belgian inquiries found none of the people, and effectively none of the addresses.

The handwriting. Kripos's analysis of her registration cards concluded she had been taught to write in France, Belgium or another francophone country. She wrote in halting German and French, and one form carried a phrase belonging to German wartime administrative vocabulary.

The teeth. Isotope analysis of her dental enamel, published by NRK in January 2018, indicated she spent her early childhood in southern or south-eastern Germany, in the region around Nuremberg, and moved west as a child or adolescent to the Lorraine area of north-eastern France, near the German and Luxembourg borders. The same work revised her likely birth year to around 1930, making her older than the 30-to-40 range originally assumed.

The DNA. Tissue preserved in formalin at Haukeland hospital since the 1970-71 autopsy yielded a full profile in 2016-2017. The Innsbruck analysis excluded Asian, Middle Eastern, African and South American ancestry and indicated European parentage. A more specific mitochondrial haplogroup result has circulated in secondary reporting but is not well sourced to a primary Kripos or university statement and should be treated with caution.

The witnesses. Hotel staff described a woman of about 1.63 m with dark hair and brown eyes, heavily made up, wearing a strong scent. She frequently changed rooms after checking in, gave her occupation variously as a travelling saleswoman and an antiques dealer, and was seen wiping surfaces. Three witnesses reported seeing her with a man in Bergen shortly before her death; all three described the encounters as silent or cold.

The Theories

What Doesn't Add Up

The case was closed on 22 December 1970, before the toxicology report arrived on 2 January 1971 and before the full autopsy report arrived on 7 January. The conclusion preceded the evidence.

The lead investigator ruled out espionage "with certainty" at a press conference. Half a century later, the same state's security files on her movements were found to exist, and former officers said the investigation had been blocked from following her abroad. Both of those things cannot comfortably be true.

She removed every label from her clothing but kept a sewing kit printed with the name and address of a Geneva hotel.

She was buried in a zinc coffin so that she could be identified later. Fifty-six years later, the coffin has done its job — there is a DNA profile, an isotope map of her childhood and a reconstruction of her face — and she is still unnamed.

Accounts conflict on several points. Reporting differs on whether her remains were exhumed after 2016 or whether only archived tissue and teeth were used. The number of aliases is usually given as eight, but some lists separate Fenella and Finella Lorck and some do not. The specific mitochondrial ancestry claim circulating in secondary sources has not been confirmed by any Norwegian institution.

Latest Developments

The last substantive verified development in the case is the June 2023 Neue Zürcher Zeitung investigation, which produced the on-record statement from the original investigator's son that he believes his father was wrong, and the commissioned forensic-psychiatric assessment that suicide is very unlikely. The NZZ also reported that her DNA profile had at that point only been run against the Interpol database, without a match, and that Norwegian legal and ethical restrictions had kept the case out of the forensic-genealogy work now used elsewhere to identify unknown remains — the technique that has resolved comparable cases in the United States and Sweden.

Nothing verified has changed since. Searches of Norwegian institutional sources — NRK, Kripos, Bergens Tidende, Bergensavisen — return no identification, no new forensic result and no official statement between 2024 and September 2026. Claims circulating online that her DNA has been matched to relatives in France are not supported by any Norwegian police, NRK or university source and should be disregarded until one appears.

She remains buried at Møllendal cemetery in Bergen. The grave is unmarked.

Sources

1. In 50-year-old unsolved death in Norway, an explosive new lead points to a Nazi-allied Swiss banker — Neue Zürcher Zeitung, 2023
2. 'Major breakthrough' in Norway's 46-year-old Isdal woman mystery — BBC News, 2017
3. Isdal Woman: The mystery death haunting Norway for 46 years — BBC News, 2017
4. The secret police file on the Penguin missile system — BBC World Service, Death in Ice Valley, 2019
5. Tenner avslører at Isdalskvinnen var eldre enn man trodde — NRK, 2018
6. DNA-guru: Isdalskvinnen var europeisk — NRK Dokumentar, 2017
7. Tennene til Isdalskvinnen gir nye spor — forskning.no / University of Oslo, 2018
8. Mysteriet om Isdalskvinnen: – Fant fullverdig DNA-profil — Bergensavisen, undated
9. – Har Isdalskvinnens DNA-profil — Bergens Tidende, undated
10. Norway makes international appeal to solve 46-year-old mystery — The Local Norway, 2017
11. Isdalskvinnen — Wikipedia (Norwegian), accessed 2026
12. Isdal Woman — Wikipedia, accessed 2026
13. Death in Ice Valley — Wikipedia, accessed 2026

Malaysia Airlines Flight 370

Seven hours of automated satellite handshakes from an aircraft that had switched everything else off

The Hook

The aircraft stopped talking at 01:21 on 8 March 2014. It kept answering.

Every hour for the next six hours, a satellite ground station in Perth sent a routine status query to the aircraft's satellite terminal, and every hour the terminal replied — not with a position, not with data, just an electronic acknowledgement that it was still powered and still there. Inmarsat had never designed those handshakes to locate anything. They existed for billing.

From the timing and frequency of those seven exchanges, engineers reconstructed an arc across the southern Indian Ocean, thousands of kilometres long, and that arc is why anyone knows the aeroplane went south at all. The final exchange, at 00:19 UTC, was a partial one — the pattern expected when both engines have run dry, the aircraft has lost main electrical power, and the auxiliary power unit has restarted the satellite terminal for a few last seconds.

Twelve years later that is still the best evidence in the case. The rest is thirty-odd pieces of debris that washed up on the far side of an ocean, and an empty seabed.

The Facts

The aircraft was a Boeing 777-200ER, registration 9M-MRO, delivered to Malaysia Airlines in May 2002. It carried 239 people: 227 passengers and 12 crew, all twelve crew Malaysian. The largest passenger group was Chinese, 153 people. Two passengers travelled on stolen Austrian and Italian passports; both were subsequently established as Iranian migrants with no connection to the disappearance.

Flight MH370 left Kuala Lumpur International Airport at 00:41 local time on 8 March 2014, bound for Beijing Capital, due at 06:30 China Standard Time.

The last ACARS data-link transmission was received at 01:07. The next scheduled transmission, around 01:37, never arrived.

At 01:19 the flight was handed off from Malaysian to Vietnamese airspace. The Malaysian government's official transcript, released at the beginning of April 2014, gives the last words from the cockpit as "Good night Malaysian three seven zero." In the days immediately after the disappearance, Malaysian officials had told reporters the last words were "All right, good night." The discrepancy was never explained.

At 01:21 the transponder stopped.

Malaysian military primary radar — which does not depend on the aircraft cooperating — tracked an unidentified return turning back across the Malay Peninsula, then northwest up the Strait of Malacca, passing near Penang and Pulau Perak. The last radar contact was at about 02:22, over the Andaman Sea.

Then the satellite handshakes: hourly, for six more hours, ending with the partial exchange at 00:19 UTC.

Timeline

• 00:41 MYT, 8 March 2014: MH370 departs Kuala Lumpur.
• 01:07: last ACARS transmission.
• 01:19: last voice contact — "Good night Malaysian three seven zero."
• 01:21: transponder ceases. Secondary radar contact lost.
• 01:21-02:22: military primary radar tracks the aircraft turning back across the peninsula and northwest up the Strait of Malacca.
• 02:25 onwards: the satellite terminal logs back on and exchanges hourly handshakes with Inmarsat 3-F1.
• 00:19 UTC (08:19 MYT): final, partial handshake, consistent with fuel exhaustion and an auxiliary-power restart.
• 01:15 UTC: a scheduled ground-to-aircraft contact receives no response.
• 8-15 March 2014: surface and subsurface search concentrates on the South China Sea along the planned route.
• 15 March 2014Malaysia announces the aircraft was deliberately diverted; the search moves to the Indian Ocean arcs.
• 29 July 2015a flaperon washes ashore on Réunion.
• 3 September 2015French authorities confirm a serial number match to 9M-MRO.
• 17 January 2017the ATSB-led underwater search, covering roughly 120,000 km² at a cost of about A$200 million, is suspended without result.
• January-June 2018: Ocean Infinity conducts a "no find, no fee" search covering tens of thousands of square kilometres further north along the arc. Nothing is found.
• 30 July 2018the Malaysian Safety Investigation Report is published.
• December 2024Malaysia agrees in principle to a new Ocean Infinity search.
• 25 March 2025the service agreement is signed. The contracted area is 15,000 km²; the fee is US$70 million, payable only on finding the aircraft.
• 28 March 2025the search is suspended after a few days, for seasonal weather and prior vessel commitments.
• 30 December 202523 January 2026: the search resumes and then ends for the season. Across both phases, 28 operational days and approximately 7,571 km² surveyed. Nothing found.
• 8 March 2026twelve years. Ocean Infinity publishes a statement summarising the result.
• 26-29 June 2026: the Malaysian Cabinet approves a twelve-month extension, 1 July 2026 to 30 June 2027, to complete the remaining 7,428.54 km² on the same terms.
• September 2026the search is paused. Operations can only be scheduled in the calm-sea window running from November 2026 to April 2027.

The Evidence

The Inmarsat data. Seven handshakes, each carrying a burst timing offset and a burst frequency offset — how long the signal took and how its frequency was shifted by relative motion. From these, analysts derived a series of rings of possible position, the last of which is the "seventh arc." Every search since March 2014 has been conducted along it. The technique had never been used this way before and was independently reviewed; the arc itself is not seriously disputed. Where on the arc the aircraft came down is the entire problem.

The military radar. Primary radar established the turn-back. This is the strongest evidence that the diversion was not a sudden catastrophic event: something or someone flew the aircraft on a series of deliberate headings for nearly an hour after communications ceased.

The debris. The 2018 Malaysian report examined 27 significant items. Three were confirmed as coming from 9M-MRO: the Réunion flaperon and two outboard flap sections. Seven more were assessed "almost certainly," eight "highly likely" and one "likely." Additional pieces have been recovered since, principally by the independent investigator Blaine Gibson working the beaches of Madagascar and Mozambique, but Malaysia's Annex 13 investigation team disbanded in 2019 and no government body has formally examined material found after that. Claims about post-2019 fragments circulate through independent channels and are not official findings.

The drift modelling. CSIRO in Australia, working with drift specialists at the University of Western Australia, modelled where debris entering the water at various points along the arc would end up. The controlling constraint was negative: debris reached Réunion, Mozambique and Madagascar, and none reached Australia. That pattern pointed to a crash site near 35°S. The area subsequently searched by Ocean Infinity was built on this work.

The flap position. Examination of the recovered flap and flaperon sections indicated the flaps were not extended. If correct, that argues against a controlled ditching and for a high-speed, uncontrolled impact — consistent with the steep descent rates that some analysts derive from the final satellite data.

The simulator. A flight simulator in the home of Captain Zaharie Ahmad Shah was examined by the FBI. Recovered fragments of deleted data included a route running south into the Indian Ocean, flown some weeks before the disappearance. Malaysia's then transport minister stated publicly that the simulator contained a very large number of stored destinations and that there was no evidence the captain had flown that particular route with intent. The official investigation did not treat it as proof of anything.

The Theories

What Doesn't Add Up

The last words were reported one way for three weeks and another way thereafter, by the same government, with no explanation offered. It is a small thing, and it is the reason a great many people stopped believing the official account of anything else.

Malaysian military radar tracked an unidentified aircraft flying back across the country in the middle of the night, and nothing was done about it at the time. The search spent its first week in the wrong ocean as a result.

The seventh arc has now been searched across more than 140,000 square kilometres by two operators using three different generations of technology, guided by drift modelling that independently pointed at the same latitude. Ocean Infinity's own assessment, published on 8 March 2026, was that the company can say with confidence the aircraft is not where they have looked. Either the wreck is in the remaining fraction of the priority area, or a premise underlying the search is wrong.

There is also a gap in the record that has nothing to do with the ocean. Malaysia's investigation team disbanded in 2019. Debris recovered since then has been assessed privately, by volunteers and engineers with no official standing. For an aircraft loss of this magnitude, that is an extraordinary state of affairs.

Sources disagree on details. The number of debris items is given as 27 in the official report and as high as sixty by independent investigators using looser criteria. The 2018 Ocean Infinity search area is quoted variously between 80,000 and 112,000 km² depending on whether overlap and transit survey are counted. And the central disagreement — whether the aircraft was flown into the water by a person or glided in unpiloted — remains unresolved between the ATSB's earlier public position and later independent analysis of the flap fragments.

Latest Developments

The current search is contracted, funded and unfinished.

In December 2024 Malaysia agreed in principle to a renewed Ocean Infinity search; the service agreement was signed on 25 March 2025, covering a 15,000 km² priority zone on a no-find, no-fee basis with a US$70 million success fee. The first phase began on 25 March 2025 and was suspended three days later for the season. The second phase ran from 30 December 2025 to 23 January 2026. Together the two phases produced 28 operational days at sea and approximately 7,571 km² of seabed surveyed — roughly half the priority area — and no trace of the aircraft.

On 8 March 2026, the twelfth anniversary, Ocean Infinity published a statement confirming the result and noting that the company has now spent 151 days at sea and mapped more than 140,000 km² of the southern Indian Ocean since 2018. Its chief executive stated that the aircraft is not where they have looked, and that the company remains committed to returning. The families' group Voice 370 used the anniversary to press the Malaysian government to extend the contract and to open talks with other operators.

On 26 June 2026 the Malaysian Cabinet approved a twelve-month extension, announced by Transport Minister Anthony Loke and running from 1 July 2026 to 30 June 2027, on identical terms, to allow the remaining 7,428.54 km² to be completed. The ministry stated that the vessel can only be scheduled into the area during the calm-sea window from November 2026 through April 2027.

As of September 2026 the search is therefore paused rather than ended. Nothing further is expected before November. If the remaining area is swept and nothing is found, the 15,000 km² zone identified by drift and satellite analysis as the most probable resting place of 9M-MRO will have been eliminated.

Sources

1. Conclusion of the search for Malaysia Airlines flight MH370 — Ocean Infinity, 2026
2. Cabinet greenlights MH370 deep-sea search extension with Ocean Infinity for another year — Malay Mail, 2026
3. Govt extends deal with Ocean Infinity to search for MH370 wreckage — The Star, 2026
4. Govt Extends Deal With Ocean Infinity For MH370 Wreckage Search — Bernama, 2026
5. Malaysia Extends Search for Missing Malaysia Airlines Flight MH370 — The Diplomat, 2026
6. 12 years on, renewed hunt for missing Malaysia Airlines flight comes up empty — NPR/AP, 2026
7. Malaysia approves a new search for MH370 wreckage in the Indian Ocean — NPR, 2025
8. Search for MH370 postponed due to weather, to resume next summer — Malay Mail, 2025
9. Assistance to Malaysian Ministry of Transport in support of missing Malaysia Airlines flight MH370 — Australian Transport Safety Bureau, ongoing
10. What debris and the Indian Ocean told drift modellers about the MH370 search area — CSIRO, 2017
11. MH370: Evidence points to Malaysia Airlines wreck being at 35 degrees south — ABC News (Australia), 2018
12. MH370 probe considered risks from batteries and fruit — FlightGlobal, 2018
13. Malaysia confirms Flight 370 pilot plotted course to Indian Ocean on flight simulator — CBC News, 2016
14. Family of MH370 captain defends his reputation — ABC News (Australia), 2014
15. Scientist claims to solve MH370 mystery, but expert urges caution — Forbes, 2024

The Springfield Three

Three women gone from a house where nothing was taken and the only damage was a broken porch light

The Hook

Everything they owned was still in the house.

Three purses on the bedroom floor. Roughly nine hundred dollars in cash inside one of them, a salon deposit Sherrill Levitt had not yet taken to the bank. Three sets of car keys. Three cars in the driveway and carport. Stacy McCall's clothes from the night before, folded. Cigarettes belonging to a chain smoker who did not go anywhere without them. A half-finished can of Coke. A television still on, showing static from a tape that had run out. A Yorkshire terrier named Cinnamon, shut inside and frantic.

The front door was unlocked. There were no broken windows, no blood, no overturned furniture, no sign that anybody in that house had fought. The single piece of physical damage anywhere on the property was the glass globe over the porch light, which had been shattered while the bulb inside it stayed intact.

Within a few hours of the discovery, that piece of evidence was swept off the porch and thrown away by a friend who did not want to walk on broken glass barefoot.

The Facts

On the afternoon of Saturday 6 June 1992, Suzanne "Suzie" Streeter, 19, and Stacy McCall, 18, graduated from Kickapoo High School at a ceremony in the Hammons Student Center on the campus of what was then Southwest Missouri State University. Streeter's mother, Sherrill Levitt, 47, a cosmetologist at a Springfield salon, attended.

The school offered an alcohol-free lock-in overnight. Both graduates skipped it and went to a series of parties instead. Their plan for the following morning was a trip to the White Water park in Branson with their friend Janelle Kirby.

Levitt was last heard from at about 11:15 p.m. on 6 June, when she spoke to a friend by telephone about painting and refinishing a chest of drawers in her bedroom. Nothing in that conversation was later described as unusual.

At about 1:50 a.m. the last party the two graduates attended was broken up by Springfield police for noise. They drove back to Kirby's house in Battlefield, a town just southwest of Springfield, at around 2:00 a.m. The house was full of Kirby's out-of-town relatives, so Streeter and McCall changed plans again and decided to sleep at Levitt's house at 1717 East Delmar Street instead. Kirby's mother, in bed at the time, heard them arranging it: Streeter telling McCall to follow her home, McCall agreeing.

The Springfield Police Department places their departure at approximately 2:15 a.m. on 7 June. The FBI's ViCAP entry on the case gives the window for the disappearance as 2:15 a.m. to 7:30 a.m.

Both cars were at the Delmar house the next morning, along with the women's clothing, jewellery, purses and keys. Used makeup wipes in the bathroom bin indicated that Streeter and McCall had reached the house and begun getting ready for bed. Levitt's bed appeared to have been slept in.

At about 9:00 a.m., when the pair failed to arrive for the Branson trip, Janelle Kirby and her boyfriend Mike Henson drove to the house. They found the front door unlocked and the house empty except for the dog. Kirby noticed the shattered porch light globe; Henson swept the glass away. While they were there, the telephone rang twice with obscene calls from an unidentified male. Kirby later recalled that Streeter had complained about prank calls at the house since moving in that spring.

Later that day Stacy McCall's mother, Janis, came to the house after her daughter failed to answer her calls and friends confirmed that neither young woman had appeared at White Water. She found the purses, the cash, the folded clothes and the agitated dog, and called police from the house telephone. While there she played back a message on the answering machine that she described as strange. It was erased in the playing — standard behaviour for the tape machines of the period. Police later said they were very interested in that message and believed it may have contained a clue. It has never been recovered.

By the time officers arrived, the scene had been walked through by an estimated ten to twenty people. Ashtrays had been emptied. Coffee cups had been washed. One responding officer described finding several people in the yard, the door standing open, and others coming and going from inside.

Timeline

• 6 June 1992, 4:00 p.m.: Streeter and McCall graduate from Kickapoo High School.
• 6 June 1992, about 11:15 p.m.: Levitt speaks to a friend by telephone. Last confirmed contact.
• 7 June 1992, about 1:50 a.m.: police break up the last party the graduates attend.
• 7 June 1992, about 2:00 a.m.: the two return to Janelle Kirby's house in Battlefield.
• 7 June 1992, about 2:15 a.m.: they leave for 1717 East Delmar. Last confirmed sighting.
• 7 June 1992, about 9:00 a.m.: Kirby and Henson find the house empty and sweep up the broken porch-light glass.
• 7 June 1992, afternoon: Janis McCall arrives, finds the purses and the dog, and calls police. Police are notified roughly sixteen hours after the women were last seen.
• 31 December 1992a man calls the America's Most Wanted hotline with information about the disappearances. The call is disconnected when the operator tries to connect him to Springfield investigators. Police say publicly that the caller appeared to have detailed knowledge and appeal for him to make contact. He never does.
• 1993investigators search forty acres in Webster County under a warrant naming the three women. Nothing is found.
• 1996KY3 investigative reporter Dennis Graves interviews Robert Craig Cox in a Texas prison. The interview is subpoenaed and presented to a grand jury convened on the case. The grand jury returns no indictment.
• June 1997a memorial bench is dedicated to the three women in the Victim's Memorial Garden at Phelps Grove Park, Springfield.
• September 1997Levitt and Streeter are declared legally dead at the request of their family. Their case files remain classified as missing persons. Janis McCall declines to have her daughter declared dead.
• 2002investigators search a concrete company site in Webster County after a tip about men there who drove a green van. Cadaver dogs indicate three spots; bones recovered are tested and found to be far too old to be the women's.
• April 2003acting on tips, investigators excavate farmland south of Cassville with backhoes. Two items are recovered — a possible blood trace and a section of a green vehicle. Laboratory testing of the blood is inconclusive.
• 2007a local writer commissions a ground-penetrating radar scan of a parking garage at Cox South Hospital after a tip that the bodies were buried in its foundations. The operator reports three distinct anomalies. Springfield police determine the lead is not credible and note the garage was not begun until September 1993, more than a year after the disappearances.
• 2021journalist Anne Roderique-Jones launches a podcast on the case.
• June 2026thirty-four years on. Springfield police continue to describe the investigation as active. A reward fund of $43,000 remains posted.

The Evidence

The porch light. The glass globe was broken; the bulb was not. That combination is hard to produce by accident and easy to produce deliberately, by someone who wanted the entrance dark. The glass was swept up and discarded before any officer saw it in place.

The answering machine message. Heard once, by one person, in a state of alarm, and erased in the act of listening. Police considered it significant and distinct from the obscene calls Kirby had taken that morning. No content from it was ever documented beyond Janis McCall's description of it as strange.

The missing clothing. Stacy McCall's shorts, shoes and purse were in the house. Her mother observed that her yellow shirt and underwear were not. If accurate, this suggests the three were taken in whatever they were wearing to sleep, or close to it.

The green van. Multiple witnesses reported seeing a van near the house around the time of the disappearance. Police publicised a description — a 1964 to 1970 Dodge panel van, light moss green, with a white front grille — and at one point parked a matching vehicle outside police headquarters to prompt memories. The van has never been identified. Investigators have generally held that removing three adults from a house would require a vehicle larger than a car.

The absence of forensic evidence. No blood, no fingerprints of value, no fibres, no DNA profile of an intruder has ever been made public. The contamination of the scene by well-meaning friends and family in the first hours is the single most consequential fact in the investigative record.

The volume of tips. Springfield police have processed upward of five thousand tips. None has produced a body, a vehicle, or a charge.

The Theories

What Doesn't Add Up

The clean house. Three people were removed from a small home without breaking anything, without leaving blood, and without waking a neighbourhood. Either the offender had complete control from the first moment, or the women had reason to go with him without resisting. Both readings imply a level of preparation that sits oddly beside the crude damage to the porch light.

The dog. Cinnamon was shut in and, by every account, extremely agitated the next morning. A Yorkshire terrier is not a silent animal. Whatever happened, it happened with a dog barking.

The sixteen hours. Police were not called until the afternoon of 7 June, and by then the house had been entered, tidied and walked through repeatedly. The one piece of physical damage at the scene had already been swept into a bin. No serious account of this case can avoid the conclusion that the investigation began at a permanent disadvantage.

The 1992 telephone call. Someone contacted the America's Most Wanted hotline on the last day of 1992 with what police characterised as detailed knowledge, and was lost to a switchboard transfer. He never called back. If that caller was genuine, the case has been solvable by one person for thirty-four years.

The absence of anything. Not a body, not a vehicle, not a piece of clothing, not a credit card transaction, not a confession that holds up. Three people did not merely vanish from a house; they vanished from the physical record entirely. That is rare in one disappearance. In three at once, it is close to unprecedented.

Latest Developments

There have been no confirmed breaks in the case in recent years. Springfield police continue to list it under cold cases and to describe the investigation as active, and the $43,000 reward for information leading to the location and prosecution of those responsible remains posted. The FBI's ViCAP missing-persons entry on the case was last revised in March 2026 and carries no new investigative detail beyond the original circumstances.

Local reporting around the thirty-third and thirty-fourth anniversaries, in June 2025 and June 2026, documented the same position: few new leads, continued public appeals, and families who have not stopped. No public statement has been made about genetic genealogy or reprocessing of any physical evidence from the house — which is consistent with the possibility that there is no biological material from the scene worth reprocessing.

Levitt and Streeter were declared legally dead in 1997. Stacy McCall has not been, at her mother's insistence.

Sources

1. Three Missing Women — Springfield Police Department, City of Springfield, Missouri (official cold case file)
2. The Springfield Three — Springfield, Missouri — Federal Bureau of Investigation, ViCAP Missing Persons, updated 2026
3. Three Missing Women: A look back at the case 27 years later — KY3, 2017/2019 (full transcript of the 25th-anniversary report, including the green van description and the Cox interview)
4. The Springfield Three: What we know about the cold case 31 years later — KY3, 2021, updated 2023
5. Ozarks Unsolved Digital Extra: Clues in the case of Springfield's 3 missing women 34 years later — KY3, 2026
6. The Springfield 3: June 7 marks 34 years since the disappearance — KY3, 2026
7. Springfield Three — Wikipedia (used as an index to primary reporting, not as a standalone source)
8. Conviction Voided — The Washington Post, 1989 (Florida Supreme Court reversal of the Zellers conviction)
9. Missing women: Theories and investigations into The Springfield Three cold case — Springfield Daily Citizen

The Man from Taured

A handmade passport from a country that never existed, in a language no philologist could ever place

The Hook

On 29 July 1960, a Member of Parliament for Honiton stood up in the House of Commons to argue that passports were a waste of everyone's time. To make his point he described a case then being tried in Tokyo.

"It is written in a language unknown and it has remained un-identified although it has been studied for a long time by philologists," Robert Mathew told the House. "The passport is stated to have been issued in Tamanrosset the capital of the independent sovereign State of Tuarid. Neither the country nor the language can be identified, although a great deal of time has been spent in the attempt."

The man had crossed borders around the world on that document without being stopped once. He was stopped in Japan only because he ran out of money and started passing bad cheques.

Everything that makes the modern legend famous — the parallel universe, the map, Andorra, the guarded hotel room, the impossible disappearance — was added afterwards, by other people, over the following fifty years. Every one of those elements can be traced to the book or the mistranslation that introduced it.

What cannot be traced is the man. Japanese police established that his name was false, his nationality was false, his biography was false and his country was invented. They never established what was true. He was sentenced as a man of no nationality, deported, and lost.

The Facts

On 24 October 1959, a 36-year-old man recorded as John Allen Kuchar Zegrus entered Japan through Haneda Airport on a flight from Taipei, accompanied by his 30-year-old wife, who was South Korean and travelling on her own valid passport. The Japanese embassy in Taipei had issued him a visa a week earlier, on 17 October.

His passport was handmade. Japanese reporting later described it as roughly the size of a weekly magazine and obviously counterfeit on sight. It bore stamps purporting to come from Japanese embassies in several East Asian countries; those were fabricated too. The document identified its holder as a citizen and roving ambassador of a state whose name appears in the records in several mutually inconsistent forms, among them "Negusi Habesi." Beneath the national stamp was a line of text in a script no linguist could identify. The Canadian newspaper The Province printed it: "Rch ubwaii ochtra negussi habessi trwap turapa." It means nothing in any language.

By December 1959 the couple could not pay for their stay. Zegrus went to the Tokyo branches of Chase Manhattan Bank and the Bank of Korea and passed forged instruments: reportedly ¥200,000 plus $140 in traveller's cheques at Chase Manhattan and a further ¥100,000 at the Bank of Korea. In January 1960 he was arrested in connection with the bank fraud, and the passport unravelled with him.

The case would ordinarily have gone to the Criminal Investigations Bureau of the Tokyo Metropolitan Police. Because Zegrus claimed diplomatic status and insisted his arrest violated diplomatic privilege, it was passed instead to the Foreign Affairs Division of the Public Security Bureau, where it was handled by Atsuyuki Sassa — then early in a career that would end with him as the first head of Japan's Cabinet Security Affairs Office. Sassa wrote about the case decades later, most fully in a chapter of his 2016 book 私を通りすぎたスパイたち (The Spies Who Passed Me By), titled "Was He a Double Agent?"

According to Sassa's account, Zegrus said he had been born in the United States, moved through Czechoslovakia and Germany to England where he finished secondary school, flown for the Royal Air Force during the Second World War and been held as a prisoner of war by the Germans, lived afterwards in Latin America, worked as an American intelligence agent in Korea, flown in Thailand and Vietnam, and finally taken service with an Arab coalition. He said he had come to Japan on a secret mission to recruit Japanese volunteers. He also told investigators, at various points, that he was an agent of the FBI and of the CIA.

When asked to locate his country on a map, he pointed at Ethiopia and said the name should properly be written in a script resembling Arabic.

Every checkable element was checked with the governments named. Nothing was confirmed. He was indicted on the basis that his nationality was unknown.

Timeline

• 17 October 1959the Japanese embassy in Taipei issues a visa on the handmade passport.
• 24 October 1959Zegrus and his wife enter Japan through Haneda Airport from Taipei.
• December 1959he passes forged cheques at the Tokyo branches of Chase Manhattan Bank and the Bank of Korea.
• January 1960he is arrested. The passport is examined and found to be counterfeit. The case is transferred to the Foreign Affairs Division of the Tokyo Metropolitan Police Public Security Bureau.
• 29 July 1960Robert Mathew MP cites the case in the House of Commons during an adjournment debate on frontier formalities, rendering the invented state as "Tuarid" and the city as "Tamanrosset" (HC Deb, 29 July 1960, c2110).
• 10 August 1960the Yomiuri Shimbun reports the sentencing — one year's imprisonment for illegal entry and fraud — and an attempted suicide in the courtroom. On being told the sentence through an interpreter, Zegrus stood, shouted that he was going to kill himself, and cut his arms with pieces of a broken glass bottle he had concealed. He was restrained by three guards and taken by ambulance to hospital. A US Foreign Broadcast Information Service daily report of the same date carries the same account from Japanese radio.
• 15 August 1960The Province of Vancouver publishes "Man With His Own Country," rendering the invented state as "Tuared."
• Through 1960–61: the Asahi Shimbun and Mainichi Shimbun continue to cover the case. Sassa records that Zegrus attempted to sue the Tokyo police Superintendent General from prison for $1 million, alleging embezzlement. The suit went nowhere.
• 22 December 1961an FBIS translation of a Japanese radio broadcast reports that the Tokyo District Court that day sentenced "John Allen K. Zegrus, a man without nationality," to one year's imprisonment for illegal entry and passing phony cheques. The Yomiuri Shimbun of the same date carries a matching headline. This directly conflicts with the August 1960 sentencing reports; the discrepancy has never been resolved in the English-language record.
• After sentence: reported as released on time already served, deported to Hong Kong. His wife was deported to South Korea. Neither has been traced afterwards.
• 1974Jacques Bergier publishes versions of the story in which the date moves to 1954 and the man ends up confined in a psychiatric hospital.
• 1981Colin Wilson and John Grant's The Directory of Possibilities prints a single sentence using, for the first time, the spelling "Taured."
• 2013–2015the fully developed legend — Andorra, the guarded hotel room, the vanishing — circulates widely online.
• 2019–2021Natanael Antonioli, Julia Louis (writing in Fortean Times, June 2021) and the Japanese researcher posting as taraiochi independently recover the Hansard record, the FBIS broadcasts, The Province article and the original Japanese newspaper coverage, connecting the legend to Zegrus.
• 2021–2022Snopes and Full Fact publish formal debunkings.

The Evidence

The Hansard record. A verbatim parliamentary transcript from 29 July 1960, while the trial was ongoing. It establishes that the passport, the unidentifiable language and the invented state are real features of a real case, not later invention. It is also the origin of the misspelling: Mathew's "Tuarid" does not appear in any Japanese source.

The FBIS daily reports. The United States Foreign Broadcast Information Service published English translations of foreign radio broadcasts. Two relevant items survive in the published volumes: one dated 10 August 1960, matching the Yomiuri report, and one from December 1961 summarising the sentence. These are contemporaneous government-compiled documents.

The Japanese newspaper coverage. Yomiuri Shimbun, 10 August 1960, headline: "The 'Mystery Man' Who Tried to Smuggle Himself Into the Country — Attempted suicide immediately after sentencing — Fictitious nationality, fluent in 14 languages." It gives the entry date, the wife's nationality, the amount of the forged cheques and the description of the passport.

Sassa's memoir. The only detailed first-hand investigative account, published in 2016 by the officer who handled the file. Untranslated, and known in English largely through summary.

The Province, 15 August 1960. Written as a light feature sympathetic to Zegrus — "a martyr to Japanese thoroughness" — and containing the invented proclamation text. It is where "Tuarid" became "Tuared."

What does not exist. There is no contemporaneous report, in any language, of a man vanishing from a guarded hotel room in Japan. There is no report of anyone pointing to Andorra. There is no report of a country called Taured before 1981. There is no 1954 case; that date enters the record with Bergier, twenty years after the fact.

The Theories

What Doesn't Add Up

The visa. A Japanese embassy in Taipei looked at a document the size of a magazine, written in gibberish, issued by a state nobody had heard of, and stamped a visa into it on 17 October 1959. That is the single most remarkable documented fact in the case, and it is the one nobody retells.

Fourteen languages. The Yomiuri described him as fluent in fourteen. Even allowing for newspaper inflation, multilingualism on that scale is not the profile of a small-time cheque forger. Whatever his life had actually been, it had covered ground.

Two sentencings. The Japanese and American records place the same sentence — one year, same charges — in August 1960 and again in December 1961. An appeal, a retrial, or a filing error would each explain it. None has been demonstrated. It is a straightforward archival question that no one has yet answered.

The escalation to diplomatic immunity. A man caught passing cheques does not usually demand release on grounds of diplomatic privilege. Either he had convinced himself, or he understood that the claim would force the case out of the criminal bureau — which is exactly what it did.

The gap. The Directory of Possibilities prints one sentence in 1981. The full legend, with the hotel and the vanishing, is circulating by 2013. Nobody has documented what happened in between. The most-repeated version of this story has no traceable author.

Latest Developments

The identification of the legend's source is recent and rests on amateur research that professional fact-checking organisations subsequently confirmed. Natanael Antonioli began posting findings in July 2019; Julia Louis, writing on the Forteana Forums and then in the June 2021 issue of Fortean Times, located the FBIS material; and in November 2020 a Japanese researcher posting as taraiochi published the 1960–61 Japanese newspaper coverage that had never been part of the English-language discussion. Snopes published its fact check in March 2021 and Full Fact in May 2022.

No new primary material has surfaced since. Nothing is known of what became of Zegrus after his deportation to Hong Kong, or of his wife after her repatriation to South Korea. His real name, his nationality and his place of birth remain unestablished, as they were when a Tokyo court sentenced him as a man without a country.

The file is closed on the parallel universe. It is still open on the man.

Sources

1. Frontier Formalities (Simplification), Volume 627, debated 29 July 1960) — Hansard, UK Parliament, 1960 (primary record of Robert Mathew's speech)
2. Frontier Formalities (Simplification) — Mr Robert Mathew, 29 July 1960 — TheyWorkForYou / mySociety (HC Deb, 29 July 1960, c2110)
3. The Mystery of the Man from Taured — Snopes, 2021 (reproduces the Hansard extract, the August 1960 newspaper feature, the December 1961 FBIS broadcast summary and a translation of the Yomiuri Shimbun report of 10 August 1960)
4. A man who arrived in Japan with a fake passport did not mysteriously vanish — Full Fact, 2022
5. John Zegrus — Wikipedia (index to the Japanese-language sources, including Atsuyuki Sassa's 2016 memoir and the Yomiuri Shimbun citations of 10 August 1960 and 22 December 1961)
6. The Man From Taured, Solved: John Zegrus And How A Real-Life Event Became An Urban Legend — The Ghost In My Machine, 2024 (detailed reconstruction of the transmission chain from Hansard to Bergier to Wilson and Grant)
7. Inside The Legend Of The Mysterious Taured Man — All That's Interesting, 2023, updated 2026
8. What really happened to the man from a 'parallel universe' who bewildered Japan — NextShark, 2023

The Roanoke Colony

One word carved on a palisade post, no distress sign beside it, and a governor never allowed to follow it

The Hook

John White came ashore on Roanoke Island on the morning of 18 August 1590, his granddaughter's third birthday. The night before, his men had anchored offshore and sung English songs across the water, hoping someone would hear. Nobody answered.

They found fresh tracks in the sand, "CRO" cut into a tree, and the settlement enclosed by a palisade that had not been there when White left, with the single word CROATOAN carved in fair capital letters into one of its posts.

They did not find a cross. That mattered more than anything else. Before White sailed in 1587 the colonists had agreed a code: carve the destination if they moved, add a cross if the move was forced.

Inside the palisade the houses had been taken down and everything portable removed. The boats were gone. Three chests White had buried were dug up and looted, but that looked like scavenging afterwards rather than violence at the time.

He read it as it was meant to be read: they had gone sixty miles south to Croatoan, to the people of Manteo. He planned to sail there the next day.

That night the Hopewell's anchor cable snapped, leaving one working anchor, and then the weather turned — the National Park Service's account has a hurricane driving the ship off the coast entirely. The crew refused the risk. White was carried back to England and never saw the Outer Banks again.

The Facts

Sir Walter Raleigh's Virginia venture ran through three attempts: a reconnaissance in 1584; a military colony under Ralph Lane on Roanoke Island in 1585, which ended in violence with the Secotan and was abandoned in 1586; and a civilian settlement chartered on 7 January 1587 as "the Cittie of Raleigh," with John White as governor.

Between 115 and 120 people signed on — the figure varies across sources and no muster roll settles it. This group was unlike Lane's: mostly middle-class Londoners, including women and children, with no organised military force. Manteo, a Croatoan man who had already made two voyages to England, sailed with them. The destination was not Roanoke Island at all but Chesapeake Bay, chosen because Roanoke had been made unsafe by Lane's war.

The fleet left England in the spring of 1587 — sources give either 26 April or 8 May — and reached the Outer Banks on 22 July. A representative of the pilot, Simon Fernandes, then ordered the sailors to put the colonists ashore on Roanoke and go no further; Fernandes wanted the rest of the season for privateering. White had no means to compel him.

So the colonists landed on a site they had not chosen, in territory where the previous English garrison had been wiped out. Within days a colonist named George Howe was killed while catching crabs alone. A retaliatory raid on the mainland village of Dasamongueponke struck Croatoan allies by mistake; Manteo repaired the relationship and was baptised as Lord of Roanoke and Dasamongueponke, the colony's only formal alliance.

On 18 August 1587 Eleanor Dare gave birth to a daughter, christened Virginia. By then the colonists had decided they could not stay. Their plan, recorded by White, was to move about fifty miles up Albemarle Sound — inland, to the mainland. They pressed White to go home and argue for relief. He left on 27 August, nine days after his granddaughter's birth; accounts differ on where he made landfall, the National Park Service recording Ireland on 16 October and other reconstructions England on 5 November.

He was then trapped by history. With the Spanish Armada mobilising, Elizabeth prohibited seaworthy ships from leaving. White got two vessels released in April 1588; their captains went hunting Spanish prizes on the way, were attacked off Morocco, lost nearly two dozen crew and all the Roanoke supplies, and limped home. After the Armada, England banned shipping again. White did not get passage until 1590, and then only as a passenger on a privateering voyage.

Timeline

• 7 January 1587Raleigh charters "the Cittie of Raleigh," John White governor.
• 22 July 1587the ships reach the Outer Banks. Fernandes's man refuses to carry the colonists further; they are put ashore on Roanoke.
• Late July–August 1587: George Howe is killed; White's raid on Dasamongueponke strikes Croatoan allies by mistake; Virginia Dare is born on 18 August.
• 27 August 1587White sails for England, leaving a colony that has already decided to move fifty miles inland.
• April–May 1588: White's relief attempt is wrecked by its own captains' privateering and an attack off Morocco.
• 12–18 August 1590: White returns aboard the Hopewell. Smoke is sighted on Roanoke and Croatoan; seven mariners drown in the landings. On 18 August the party finds "CRO" on a tree and "CROATOAN" on a palisade post, no cross, the houses dismantled and the boats gone.
• 19 August 1590the anchor cable parts and the weather turns. The voyage to Croatoan is abandoned. White reaches England on 24 October 1590 and never returns.
• 1602–1603three further expeditions are mounted, nominally to search. None reaches Croatoan; Raleigh later admitted one was cover for a sassafras venture.
• 1607 onwards: Jamestown colonists collect Powhatan reports of places where men wore European clothing. The "Zúñiga Map," a copy of a lost Jamestown map rediscovered in 1890, notes: "Here remaineth four men clothed that came from Roonock to Ocanahawan."
• 1830s onwards: the episode is reframed as a romantic mystery — Andrew Lawler dates the phrase "Lost Colony" to this period — and the outdoor drama The Lost Colony, running at Fort Raleigh since 1937, fixes the vanishing version in public memory.
• 2012backlighting of John White's map La Virginea Pars at the British Museum reveals a fort symbol hidden under a patch, pointing to a site at the head of Albemarle Sound.
• 2015–2020First Colony Foundation excavations there — "Site X," and a second site near Salmon Creek in Bertie County — recover English, German, French and Spanish ceramics. In October 2020 the foundation announces the finds and proposes that the colony split.
• June 2020Scott Dawson of the Croatoan Archaeological Society publishes The Lost Colony and Hatteras Island, arguing all the colonists went to Croatoan.
• March 2024excavation at the Elizabethan Gardens recovers 16th-century Algonquian pottery and a drawn-copper ornament, placing the Native village of Roanoac.
• Summer 2024excavation on Hatteras Island recovers hammerscale — blacksmithing waste — from a pit beneath a shell midden containing almost no European material. It is reported publicly in June 2025 and has not been peer reviewed.
• 26 March 2026at the Coastal Studies Institute in Wanchese, Mark Horton and Beth Scaffidi present what they call a potential "smoking gun" — a lead-and-iron ball of Elizabethan composition embedded in a deer jawbone from Hatteras, alongside hammerscale and bar iron.

The Evidence

The carvings. White's account describes CROATOAN in "fayre Capitall letters" on a palisade post and CRO on a tree, with no cross at either. The carvings are long gone; the evidence is White's narrative, preserved in Hakluyt.

The dismantled settlement. Houses taken apart rather than burned, everything portable removed, boats gone — the signature of a planned move, not a massacre. The looted chests are the one discordant note, and they were buried items anyone could have found later.

The colonists' stated plan. Before White sailed, the settlers intended to move fifty miles up Albemarle Sound, inland. That is not Croatoan.

The White map and the inland ceramics. La Virginea Pars, at the British Museum, carries a fort symbol concealed beneath an applied patch, revealed by backlighting in 2012. It sits at the head of Albemarle Sound, close to where First Colony Foundation later recovered English, German, French and Spanish vessel fragments from two Bertie County sites. Nick Luccketti called the quantity and variety "compelling evidence" of 1587 settlers. Critics, Scott Dawson among them, argue that European pottery inland is exactly what trade produces.

The hammerscale. Flakes of iron oxide struck off during forging. It is waste, not merchandise, so it does not travel through trade the way a pot does, and iron forging was not part of Indigenous practice in the region. It was found in a pit sealed beneath a shell midden containing almost no European material, in a layer radiocarbon-dated consistent with the Lost Colony period.

The musket ball in the deer. Recovered from Hatteras, initially mistaken for an arrowhead, embedded in a deer's jaw; compositional testing found lead and iron in proportions consistent with Elizabethan ammunition. Beth Scaffidi's formulation is careful: "we can't say who shot it. But we can say that ammo would not be in that deer but for a European person showing up on Hatteras Island in the late 16th or early 17th century."

The faunal isotope work. In the absence of human remains, the team reads discarded animal bone from village middens for stable isotopes that track diet. Early results indicate Indigenous and European food systems blending, during the coldest and least productive stretch of the Little Ice Age.

The Jamestown hearsay. Powhatan accounts of clothed men at inland settlements, and the Zúñiga Map note — second-hand, through interpreters, two decades later. Suggestive, not proof.

The Theories

What Doesn't Add Up

Sixty miles. White never sailed them. One parted cable in August 1590 is the reason this case exists. Had the Hopewell reached Croatoan there would in all likelihood be no mystery and no four centuries of theorising.

Nobody searched afterwards. Raleigh's Virginia patent held only so long as the colonists could not be proven dead, so he had a legal interest in their remaining unaccounted for. Expeditions were funded with a stated search purpose; one was admitted afterwards to have been cover for sassafras. The people best placed to find them had a reason not to.

The plan and the message disagree. The colonists told White they were going fifty miles up Albemarle Sound. They left a word pointing sixty miles south to an island. Either the plan changed, or the message was meant for part of the group only, or it names a staging point rather than a destination.

Fresh tracks. White's party found footprints in the sand on 18 August 1590 and saw smoke on both Roanoke and Croatoan in the preceding days. Somebody was there. Nobody made contact, and White did not use his days anchored off Croatoan to send for the Croatan.

No graves. Ewen's test is the right one and has never been met: European burials, securely dated to the 16th century, with European material that is not trade goods. Four centuries of searching has not produced a single Lost Colonist's body.

Latest Developments

Two research programmes are working the case, and they disagree.

On Hatteras Island, Mark Horton of the Royal Agricultural University and the Croatoan Archaeological Society have been excavating since around 2009. The 2024 hammerscale, Horton's "real breakthrough," was followed in March 2026 by bar iron and the musket ball in the deer jaw. Beth Scaffidi of UC Merced is running isotope and radiocarbon analysis on faunal middens to reconstruct diet and land use, with deer bones showing arrow wounds and, in the 17th century, gunshot wounds. Horton characterises the result as "peaceful assimilation": a survivors' camp that dissolved into the local villages as food ran out. The team acknowledges small sample sizes, radiocarbon margins that blur the crucial fifty years either side of 1590, and erosion damaging the sites. Excavation is largely complete and publication is said to be next. Until then the strongest claim in the case rests on a lecture rather than a paper.

Inland, the First Colony Foundation continues to work the Albemarle Sound sites, and in March 2024 returned to the Elizabethan Gardens, where Eric Klingelhofer's team located the Algonquian village of Roanoac itself. Compositional testing has since supported the reading that the copper ornament found there was traded or given to Native people by Raleigh's explorers — a 1584 question rather than a 1587 one, but proof that the material can answer this class of question.

Each camp reads the other's evidence as trade. That is the live problem: at what point does European material at a Native site stop being commerce and start being cohabitation. Neither team has produced the thing that would end the argument, which is a grave.

Sources

1. CROATOAN — Phillip W. Evans, Encyclopedia of North Carolina (University of North Carolina Press, 2006), via NCpedia
2. Origins of the "Lost Colony" Mystery — North Carolina Department of Natural and Cultural Resources, 2016
3. Roanoke Colony — Wikipedia (used as an index to the primary narrative in Hakluyt and to the scholarship of David Beers Quinn, Karen Kupperman and Andrew Lawler)
4. New evidence from Hatteras reshapes the Lost Colony story — Brian Tress, The Outer Banks Voice, March 2026 (report on the Coastal Studies Institute presentation of 26 March 2026)
5. 'Lost Colony' of Roanoke may have assimilated into Indigenous society, archaeologist claims — but not everyone is convinced — Kristina Killgrove, Live Science, June 2025
6. Pottery Fragments May Hold Clues to Roanoke Colonists' Fate — Livia Gershon, Smithsonian Magazine, 2020
7. Recent findings shed light on the "Lost Colony" of Roanoke — Mark Milligan, HeritageDaily, May 2024 (First Colony Foundation excavations at the Elizabethan Gardens)
8. John White — National Park Service, Fort Raleigh National Historic Site
9. Lost Colony News & Research — First Colony Foundation

The Methuselah Star

A star 190 light-years away has been measured, repeatedly, as older than the universe that contains it

The Hook

In February 2013, Howard Bond of the Space Telescope Science Institute and four co-authors published a four-page letter in the Astrophysical Journal. Its subject was HD 140283, a seventh-magnitude star in Libra that is bright enough to find in binoculars and has been catalogued for more than a century. Using eleven epochs of astrometry from the Fine Guidance Sensors on the Hubble Space Telescope, they had measured its distance five times more precisely than the Hipparcos satellite had. Fed into stellar isochrones, that distance returned an age of 14.46 billion years.

The best measurement of the age of the universe available to them, from the cosmic microwave background, was 13.77 billion years, with an uncertainty of 60 million. The star was apparently 700 million years older than everything.

Bond's paper did not claim a paradox. It stated the opposite: that once the uncertainty in the star's oxygen abundance was folded in, the total error reached ±0.8 billion years, and the star's age "does not conflict with the age of the Universe." That nuance did not survive contact with the press. The object was nicknamed the Methuselah star, and it has spent thirteen years as the standing example of a measurement that appears to break cosmology.

The interesting part is what happened afterwards. The number has been recalculated at least eight times since, driven down to 12.0 billion years and back up to 14.2. Every one of those revisions was legitimate. None of them settled the question.

The Facts

HD 140283 is a Population II subgiant: a low-mass star that has exhausted the hydrogen in its core and begun to expand, but has not yet reached the red giant branch. Its surface iron content is about one five-hundredth of the Sun's, expressed by astronomers as [Fe/H] = -2.3. Its effective temperature is close to 5,790 K, almost the same as the Sun's, and its radius is roughly 2.1 times solar.

Three properties make it the best age target in the sky. It is bright, so its spectrum can be measured in detail. It sits behind no measurable interstellar dust, removing the reddening correction that plagues globular cluster work. And subgiants of this kind occupy the one part of the Hertzsprung-Russell diagram where absolute brightness is most sensitive to age.

The star also has a place in the history of the field. Walter Adams identified it in 1912 as a high-velocity object with a spectral type too early for its luminosity. In 1951 Joseph Chamberlain and Lawrence Aller showed, using HD 140283 and the similar star HD 19445, that some stars simply contain far less heavy-element material than the Sun — a result that fed into the 1957 synthesis of Burbidge, Burbidge, Fowler and Hoyle establishing that heavy elements are made inside stars. Low metal content became the signature of great age, and HD 140283 was the first object to carry it.

Its motion is extreme. The Hubble astrometry gave a tangential velocity of 319 km/s and a total space velocity relative to the Sun of 361 km/s, on a highly eccentric, retrograde orbit through the Galactic halo. This is not a disc star.

The central number, the age, depends on four inputs: the distance, the surface temperature, the oxygen abundance, and the efficiency of convection in the star's outer envelope. Only the first two are measured directly. The other two are the argument.

Timeline

• 1912Walter Adams notes HD 140283 as the first high-velocity star with an anomalously early spectral type for its luminosity.
• 1951Chamberlain and Aller demonstrate spectroscopically that HD 140283 and HD 19445 are severely deficient in heavy elements.
• 1968Cohen and Strom, and independently Cayrel, establish that the star is a slightly evolved subgiant rather than a main-sequence subdwarf, placing it at the most age-sensitive point of the HR diagram.
• 1997The Hipparcos catalogue gives a parallax of 17.16 ± 0.68 milliarcseconds. Isochrones without element diffusion imply an age above roughly 14 billion years. Don VandenBerg's 2000 analysis implies that if the star is used to calibrate globular clusters, the oldest clusters are at least 15 billion years old.
• 2002VandenBerg, Richard, Michaud and Richer include helium diffusion and reduce the star's age to 13.5 ± 1.5 billion years.
• 2004Revision of the 14N(p,γ)15O nuclear reaction rate pushes model ages back up by 0.7 to 1.0 billion years.
• 2003-2011Bond's team observes the star at eleven epochs with Hubble's Fine Guidance Sensors.
• February 2013Bond, Nelan, VandenBerg, Schaefer and Harmer report a parallax of 17.15 ± 0.14 mas and an age of 14.46 ± 0.31 billion years, with a total uncertainty of ±0.8 billion. The "oldest star" story goes worldwide.
• 2014VandenBerg, Bond, Nelan and colleagues publish a companion study of three ancient halo subgiants and obtain 14.27 ± 0.38 billion years.
• 2015Orlagh Creevey's team, using interferometry from the CHARA Array, reports 13.7 ± 0.7 billion years — the first result comfortably inside the age of the universe.
• 2018Meridith Joyce and Brian Chaboyer find a mass-dependent age range of 12.5 to 14.9 billion years, the spread driven by the treatment of convection. Karovicova and colleagues measure an angular diameter of 0.324 ± 0.005 mas and a temperature of 5,787 ± 48 K from CHARA.
• 2020Karovicova's team publishes a homogeneous set of fundamental parameters for ten metal-poor benchmark stars, combining CHARA interferometry with Gaia astrometry. Gaia's parallax for HD 140283, 16.26 ± 0.026 mas in Data Release 3, is measurably smaller than both the Hipparcos and Hubble values.
• 2021Jianling Tang and Meridith Joyce, using MESA models and the new CHARA radius, report 0.81 ± 0.05 solar masses and an age of 12 ± 0.5 billion years.
• November 2024Guillaume, Buldgen, Amarsi and collaborators compute opacity tables tailored to the star's own measured abundances and report 12.3 billion years; the same models with a generic solar-scaled composition give 14.1 billion.
• 13 October 2025Lundkvist, Larsen and a fourteen-author team post the first asteroseismic analysis, using oscillations detected by TESS. They report a mass of 0.75 ± 0.01 solar masses and an age of 14.2 ± 0.4 billion years.

The Evidence

The physical measurements are now very good, and they are not the problem.

The distance has been measured three times. Hipparcos gave a parallax of 17.16 ± 0.68 mas; Hubble's Fine Guidance Sensors gave 17.15 ± 0.14 mas, agreeing closely. Gaia DR3 gives 16.26 ± 0.026 mas, which does not agree: it is roughly 0.9 mas lower, well outside the Hubble error bar, and places the star about three parsecs farther away. Gaia's formal precision is the highest of the three by an order of magnitude, and most recent work adopts it. The disagreement between two high-precision astrometric missions on a bright, nearby, well-behaved star has not been publicly resolved.

The size has been measured directly. CHARA, an optical interferometer on Mount Wilson, resolved the star's disc at 0.324 ± 0.005 milliarcseconds. Combined with the bolometric flux this gives an effective temperature of 5,787 ± 48 K without recourse to models.

Since October 2025 there is a fourth, independent handle. TESS observed HD 140283 at 20-second cadence in Sector 51, and Lundkvist's team extracted a custom light curve from the target pixel files after finding the standard pipeline products compromised by Earth and Moon glints in Camera 1. The power spectrum contains solar-like oscillations, including mixed dipole modes that probe the stellar core directly. This is the first time the interior of the star has been sounded rather than inferred. The modes give a mass of 0.75 ± 0.01 solar masses.

What has never been measured well is the oxygen abundance. Bond's own error budget is explicit: of a total age uncertainty of 0.80 billion years, 0.61 came from the ±0.15 dex uncertainty in [O/H] alone, more than parallax, photometry, extinction and iron abundance combined. Oxygen can be measured three ways in a star like this — a forbidden line at 6300 Å, a permitted triplet near 7772 Å, and molecular OH bands — and all three are unreliable, being respectively too weak, distorted by departures from local thermodynamic equilibrium, and contaminated by granulation. Published values for HD 140283 spanned [O/H] = -1.55 to -1.78; Bond took the mean. Adding 0.30 dex of oxygen would have dropped the age to 13.3 billion years on its own.

The second soft input is the mixing-length parameter, a single number standing in for the entire physics of convection in the star's envelope. Tang and Joyce found their models preferred a value 10 to 20 per cent below the solar-calibrated one, which lowers the age. Guillaume's team found no preference either way. Lundkvist's asteroseismology, which constrains it most directly, returned 1.80 ± 0.05 — consistent with the solar value, and therefore no help in lowering the age.

The Theories

What Doesn't Add Up

The parallax disagreement is the loose thread. Hubble's Fine Guidance Sensors and Gaia are both instruments designed for exactly this task, and on this star they differ by roughly six times Hubble's quoted uncertainty. Bond's team cross-checked their proper motion against Hipparcos and matched it to within a milliarcsecond per year, so the solution is not obviously broken. HD 140283 is also bright, and bright stars are a known difficulty for Gaia. Neither side has published a retraction, and most authors simply use the newer number.

The 2025 asteroseismic result also sits awkwardly. It was expected to end the argument, and in one sense it did — it delivered a mass to one per cent, 0.75 solar masses, where classical fits had scattered between 0.75 and 0.81. But it returned an age of 14.2 billion years, which is where Bond and VandenBerg were in 2013 and 2014, and 0.8 sigma above the Planck age of the universe. The authors are careful to say this is consistent within one sigma and that their quoted uncertainty covers internal errors only, with realistic systematics for a subgiant perhaps nearer 10 per cent. Still, the most direct probe yet made of the star's interior moved the number back up, not down. In the same paper, the seismic radius of 2.078 solar radii comes out 1.5 sigma smaller than CHARA's interferometric value — a known bias, but it means the two best direct measurements of the star's size do not quite agree either.

Finally, the star may not be from here. Lundkvist's team integrated its orbit and found it highly eccentric and retrograde, sitting in the region of velocity and energy space associated with Gaia-Enceladus, a dwarf galaxy the Milky Way absorbed 8 to 11 billion years ago. If that is correct, HD 140283 formed somewhere else entirely and was delivered here by a collision.

Latest Developments

The asteroseismic analysis by Lundkvist and colleagues was submitted in July 2025, accepted on 30 September and posted on 13 October 2025, appearing in Astronomy & Astrophysics the following month. It is the current state of the case: mass 0.75 ± 0.01 solar masses, radius 2.078 solar radii, age 14.2 ± 0.4 billion years, mixing-length parameter consistent with solar.

Two follow-ups are already identified by the authors themselves. TESS observed the star again in Sector 91, doubling the available 20-second-cadence data; that sector was not included in the 2025 analysis and may yield additional oscillation modes. More importantly, the asteroseismic modelling used standard abundances and opacities rather than the tailored tables Guillaume's group built in 2024 — the same substitution that moved the classical age by nearly two billion years. Combining the two approaches has not yet been published.

The team also reported that the star's frequency of maximum oscillation power, 611.3 ± 7.4 microhertz, is 14 per cent higher than the standard scaling relation predicts — a reminder that the newest tool being used to date HD 140283 is itself not fully calibrated at this composition.

The case stays open for an unglamorous reason. No measurement has yet been made that fixes the age independently of a modelling choice. The distance is disputed between two spacecraft. The oxygen abundance is uncertain at the level that dominates the answer. The convective efficiency is degenerate with the composition. The star is old — every method agrees, and nothing plausible makes it younger than about 12 billion years — but the last billion years, the ones that decide whether there is a paradox at all, are still argued from assumptions rather than read off an instrument.

Sources

1. HD 140283: A Star in the Solar Neighborhood that Formed Shortly After the Big Bang — Bond et al., Astrophysical Journal Letters 765, L12, via NASA/STScI, 2013
2. HD 140283: A Star in the Solar Neighborhood that Formed Shortly after the Big Bang (abstract) — NASA Astrophysics Data System, 2013
3. Asteroseismic investigation of HD 140283: The Methuselah star — Lundkvist et al., arXiv preprint of Astronomy & Astrophysics, 2025
4. Asteroseismic investigation of HD 140283: The Methuselah star (journal version) — Astronomy & Astrophysics, 2025
5. The age of the Methuselah star in light of stellar evolution models with tailored abundances — Guillaume et al., arXiv preprint of Astronomy & Astrophysics, 2024
6. The age of the Methuselah star in the light of stellar evolution models with tailored abundances (journal version) — Astronomy & Astrophysics, 2024
7. Revised Best Estimates for the Age and Mass of the Methuselah Star HD 140283 using MESA and Interferometry — Tang & Joyce, Research Notes of the AAS, 2021
8. Accurate effective temperatures of the metal-poor benchmark stars HD 140283, HD 122563 and HD 103095 from CHARA interferometry — Karovicova et al., MNRAS Letters, 2018
9. Fundamental stellar parameters of benchmark stars from CHARA interferometry. I. Metal-poor stars — Karovicova et al., Astronomy & Astrophysics 640, A25, 2020
10. Benchmark stars for Gaia: fundamental properties of the Population II star HD 140283 — Creevey et al., Astronomy & Astrophysics, 2015
11. Three Ancient Halo Subgiants: Precise Parallaxes, Compositions, Ages, and Implications for Globular Clusters — VandenBerg et al., Astrophysical Journal 792, 110, 2014
12. High-resolution abundance analysis of HD 140283 — Siqueira-Mello et al., Astronomy & Astrophysics, 2015
13. Constraining the age of HD 140283 — Uppsala University thesis, DiVA portal, 2022
14. Methuselah: The oldest star in the universe — Space.com

The Lost Cosmonauts

Two Italian brothers recorded a woman's voice they said was burning up in orbit; no such flight has ever been found

The Hook

The recording is two minutes and thirty-three seconds long. A woman's voice, buried in hiss, speaking Russian. In the translation published for sixty years she says: I am hot. I am hot. Is this dangerous? I can see a flame. Thirty-two. Forty-one. Am I going to crash? I will reenter. I am listening. I am hot.

The men who recorded it were two brothers from Turin, Achille and Gian Battista Judica-Cordiglia. They said they had picked it up in May 1961 from a Soviet spacecraft on a mission that never officially existed, carrying a cosmonaut whose death was being erased as it happened. They had, they said, already recorded a Morse SOS from a stranded capsule in November 1960 and the heartbeat and laboured breathing of a man dying in orbit in February 1961.

No Soviet launch corresponding to any of these events has ever been found — not in the archives opened after 1991, not in the memoirs of the engineers, not in the recollections of the cosmonaut corps. Every serious historian of Soviet spaceflight has rejected the claims.

And yet the case has never quite closed, for one reason: the Soviet space programme really did hide its dead. A trainee burned to death in a pressure chamber three weeks before Gagarin flew, and the country was not told for twenty-five years. The question the recordings pose is not whether the USSR concealed fatalities. It is whether two amateurs outside Turin, with surplus American military receivers and a hand-cranked antenna, heard one.

The Facts

Achille Judica-Cordiglia (1933-2015) and Gian Battista Judica-Cordiglia (1939-2024) took up amateur radio in 1949. In 1959 the family moved to Turin, and the brothers set up a listening post in a disused wartime bunker at San Maurizio Canavese, which they called Torre Bert. They tracked Sputnik 1 on a US war-surplus receiver, then Sputnik 2 and 3, and equipment was later donated by the Geloso radio company. They built a steerable antenna array, wall displays showing satellite orbits, and a network of sixteen correspondent stations they called Zeus. Their sister Maria Teresa was assigned to learn Russian and translate intercepts.

Sven Grahn, a Swedish space engineer who examined photographs of the installation in detail, concluded that the station was "reasonably well equipped" and that the brothers genuinely did receive signals from spacecraft. The main antenna's azimuth appears to have been steered by hand, using a wheel — a serious limitation when a low-orbiting satellite crosses the sky in minutes.

The sensational intercepts, as described in the brothers' own two-issue bulletin Radiospazio (May and June 1963), are these:

• 28 November 1960: a Morse message, "SOS to the whole world," from a spacecraft leaving Earth on an escape trajectory.
• 2 February 1961, at 22:55: heartbeats and heavy breathing from a dying cosmonaut.
• 16-23 May 1961 (dates vary in different tellings): the female voice recording.
• 14 June 1963: a female voice which they said proved Valentina Tereshkova had been launched two days earlier than announced.

The May 1961 recording did not reach the public until the Italian newspaper Corriere della Sera ran it in March 1965, and in that first telling the flight was dated to November 1963, not 1961. The date was revised later.

Torre Bert closed in 1966; the Zeus network lasted about a year. Both brothers went on to conventional careers, Achille in cardiology and Gian Battista in forensic medicine.

Timeline

• December 1959The Italian news agency Continentale reports, sourcing an unnamed senior Czechoslovak communist, that Soviet cosmonauts have already died on suborbital flights, naming Alexei Ledovsky, Serenty Schriborin, Andrei Mitkov and Maria Gromova. Independently, the rocket pioneer Hermann Oberth claims a pilot was killed on a ballistic flight from Kapustin Yar in early 1958, citing Italian press reports.
• 1959The Soviet magazine Ogoniok publishes photographs of three high-altitude parachutists — Pyotr Dolgov, Ivan Kachur and Alexey Grachov. Their names will later appear on Western lists of dead cosmonauts.
• 15 May 1960Korabl-Sputnik 1 launches. Its retrorockets fire with the craft wrongly oriented, pushing it into a higher orbit instead of bringing it down. It carried no crew and no heat shield. Robert Heinlein, in Vilnius that day, is told by Red Army cadets that a man has been launched; the report is denied hours later.
• 24 October 1960The Nedelin catastrophe at Baikonur kills a large number of personnel when an R-16 second stage ignites on the pad. It is concealed for three decades.
• 28 November 1960Torre Bert claims the Morse SOS intercept.
• 2 February 1961Torre Bert claims the heartbeat and breathing intercept.
• 23 March 1961Valentin Bondarenko, 24, a member of the first cosmonaut group, dies of burns after an alcohol-soaked swab ignites on a hot plate in an oxygen-rich low-pressure chamber at the Institute of Biomedical Problems in Moscow. His death is not disclosed.
• 10 April 1961Dennis Ogden writes in the British Daily Worker that the announcement of a car crash involving test pilot Vladimir Ilyushin is cover for a failed orbital flight on 7 April.
• 12 April 1961Yuri Gagarin orbits the Earth. The brothers say they recorded his voice on the 20 MHz band after a tip from the Italian state press agency.
• 16-23 May 1961: The claimed date of the female-voice intercept.
• 1963After the New York Journal American lists the parachutists among the dead, Izvestia and Krasnaya Zvezda publish a rebuttal with photographs and testimony from the living men. The parachutists write to the paper's editor; the letter is ignored.
• 1964The US edition of the Guinness Book of World Records prints a footnote naming nine putative lost cosmonauts.
• March 1965Corriere della Sera publishes the female-voice story.
• 1965 and 1971: Julius Epstein's papers on "Soviet failures in space" are read into the US Congressional Record.
• 1966Torre Bert shuts down.
• 1975James Oberg publishes "Phantoms of Space," systematically dismantling the phantom-cosmonaut lists.
• 1986Yaroslav Golovanov, interviewing the retired parachutist Alexey Belokonov, establishes that the men on the lists never flew to space. In April, Izvestia discloses Bondarenko's death for the first time, twenty-five years after the fact.
• 1991 onwards: Soviet archives open. The N1 lunar programme, the Nedelin disaster and the full record of the cosmonaut corps become public. No unaccounted flight appears.
• 13 November 2025David Chappell publishes a word-by-word transcription of the female-voice recording, concluding that it contains no reference to spaceflight whatsoever.

The Evidence

The physical evidence is four audio files and a set of claims about frequencies. The counter-evidence is structural.

Grahn's technical objections are specific. The brothers said they had received facsimile images from Kosmos satellites on 145.800 MHz, a band used only by amateur satellites, and identified a US long-range bomber in one of them; the early Zenit reconnaissance satellites that did send facsimile images used 163 MHz and produced pictures of poor quality. They said they had received lunar surface images from Luna 4 — but no comparable reception was logged by Jodrell Bank or the US Naval Research Laboratory's 45-metre dish, and the E-6 lunar probes are not known to have transmitted images during approach. They described a spacecraft flung onto an escape trajectory by a misfiring retrorocket, which would require a velocity change of roughly 3.2 km/s where a real retrorocket delivers around 100 m/s — a vehicle carrying more than twice its dry mass in propellant, far beyond the Vostok booster.

The heartbeat recording has a cleaner objection. Soviet biomedical data from Vostok and Voskhod was never sent over the voice channel. Heart rate and respiration were encoded as telemetry, which on an audio receiver sounds like noise. Grahn's question is the one that matters: why would a failed flight use a different transmission system from the successful ones?

Then there is the negative evidence, which is the strongest of all. Between 1991 and the present, the early Soviet crewed programme has been comprehensively declassified. The composition of the first cosmonaut group is known — who flew, who was dismissed, who died in ground tests. Engineers, cosmonauts and their families have published memoirs. Mark Wade, of the Encyclopedia Astronautica, has stated flatly that Vladimir Ilyushin was never a cosmonaut. Ilyushin himself lived until 2010 and never supported the claim. Not one of the phantom missions has produced an engineer who worked on it, a relative who mourned someone, or a document.

The newest evidence concerns the recording itself. In November 2025 David Chappell published a transcription made by filtering the audio in Audacity and working utterance by utterance rather than listening straight through. His version differs from the circulated translation at nearly every critical point. Where the old translation has "oxygen, oxygen," the speaker says dykhaniye — "breathing." Where it has the anxious "isn't this dangerous?" she says "nothing dangerous" and then "all done," closing the subject. Where it has "am I going to crash?" she says "hello, hello — oh yeah." The translation omits her saying "he is listening," which implies a third person standing beside her in a room. Between businesslike exchanges she plays with the rhyme between lekarstva (medicine) and mne zharko (I'm hot), and asks for medicine for an imaginary fever.

Chappell's conclusion is that the woman is a radio operator in a hot room, dictating and receiving message traffic, and amusing herself. "I see a flame" is real and is one of the clearer phrases in the tape — but it is delivered with a two-second pause for effect, confuses the operator at the other end, and is bracketed by nonsense.

The Theories

What Doesn't Add Up

The dates move. The female-voice flight was first reported in 1965 as a November 1963 mission. It was later redated to 16-19 May 1961, and elsewhere to 19-23 May. A recording made on a specific evening in a logged listening session should have one date. The Radiospazio bulletins themselves give reception frequencies only as "around 20 MHz."

The heartbeat tape is the claim least explained by the misinterpretation theory. Ambient noise and a bored operator do not produce a pulse. But nothing in the Vostok system transmitted a pulse over voice either, which leaves the recording without a plausible source on either reading.

The Gagarin recording raises its own question. The brothers say they were tipped off and listened on the 19.995-20.005 MHz band; Vostok 1's voice link was at 20.006 MHz, so the interception is technically possible. Grahn, watching a 2008 documentary, noted that the audio played sounded very much like the official Soviet recording released at the time on a commercial EP. He does not assert substitution. He says it made him suspicious, and that he would like to hear the brothers' unedited original.

And Chappell raises a small, uncomfortable point about the female-voice tape. The unheard operator at the other end is given almost no time to speak. The pattern is consistent across the recording. He suggests the tape may have been edited to remove silences — which would be an ordinary thing to do to a two-and-a-half-minute excerpt of a long session, and also the sort of edit that turns disconnected remarks into what sounds like a single escalating emergency.

Latest Developments

The substantive development is Chappell's transcription, published on 13 November 2025 — the first systematic attempt anyone has made to establish what the recording actually says rather than whether the event it supposedly describes occurred. His finding reframes the case: the argument for six decades has been "real distress call versus clumsy fake," and his answer is neither. The words are real, the Russian is native, and the subject is not space.

That leaves the Judica-Cordiglia brothers in an unusual position. They are no longer the most likely suspects in a hoax. They are two enthusiasts whose expectations, and an unqualified translation, converted a bored radio operator's complaints about the heat into a woman dying in orbit — and who then spent the rest of their lives defending it. Gian Battista gave interviews maintaining the claims into the 2020s. He died in 2024, and Achille in 2015, so neither can now be asked about the original tapes, the logs, or the frequencies.

The case remains open rather than closed because the primary material has never been examined the way the voice recording finally was. The Morse SOS of November 1960 and the February 1961 heartbeat tape have not received an equivalent forensic treatment. The original reels, the station logs and the raw frequency records — the things Grahn asked for repeatedly and never received — have not been made public. What exists is a set of copies of copies, a claim, and an absence in every archive that ought to contain the corroboration.

Sources

1. Notes on the space tracking activities and sensational claims made by the Judica-Cordiglia brothers — Sven Grahn, Space Tracking Notes, with postscript 2008
2. Words Without Burden: The Cosmonaut Ludmila Recording And Its Misinterpretation — David Chappell, Eclectic, 2025
3. Words Without Burden (PDF) — David Chappell, 2025
4. Lost Cosmonaut Rumors — Jeff Scott, Aerospaceweb.org, 2005
5. Lost Cosmonauts — Wikipedia, citing Oberg 1988, Golovanov 1986, Siddiqi 2003
6. Judica-Cordiglia brothers — Wikipedia
7. Chapter 10: Dead Cosmonauts, from Uncovering Soviet Disasters — James Oberg, Random House, 1988
8. Phantoms of Space — James Oberg, via Encyclopedia Astronautica, 1975
9. Valentin Bondarenko — Wikipedia
10. Soviet cosmonaut death reported 25 years later — UPI Archives, 1986
11. The last hours of Valentin Vasilyevich Bondarenko — The Orbiter, 2025
12. Bondarenko — Encyclopedia Astronautica
13. Search for the Missing Cosmonauts (Skeptoid #115) — Brian Dunning, Skeptoid, 2008
14. The Lost Cosmonauts: A Story of Soviet Secrets and Space Hackers — The Debrief
15. Judica-Cordiglia brothers — HFUnderground wiki

The Green Children of Woolpit

Two 12th-century English chroniclers, writing independently, record a brother and sister with green skin who would eat only raw beans

The Hook

There is no body, no artefact, no parish record. The entire case consists of two passages of Latin prose written decades apart by men who lived two hundred and fifty miles from each other and, as far as anyone can determine, did not read each other's work.

William of Newburgh, an Augustinian canon in Yorkshire, put it in his Historia rerum Anglicarum around 1189. Ralph of Coggeshall, the Cistercian abbot of a house in Essex, put it in his Chronicon Anglicanum in the 1220s. Both describe the same event: at harvest time, in the Suffolk village of Woolpit, villagers came upon a boy and a girl whose skin was green. Their clothing was of unfamiliar material. Their language was unintelligible. They wept and refused everything offered to them until someone brought raw broad beans, which they ate ravenously.

The boy weakened and died around the time the two were baptised. The girl lived, lost her colour, learned English, and gave an account of where they had come from: a country where the sun never rose, where the light was permanent twilight.

William was uneasy about including it. His own phrasing — sum obrutus ut cogerer credere — reads literally as "I am crushed sufficiently that I am forced to believe it." He was a chronicler who elsewhere called Geoffrey of Monmouth's history "gushing and untrammeled lying." He put the green children in anyway.

The Facts

Woolpit is a real village, recorded in the Domesday Book in 1086 as a substantial settlement in the hundred of Thedwastre, held by the Abbey of Bury St Edmunds. Its name comes from the Old English for a wolf pit: a trap dug into the ground, baited with carrion and covered with branches. In the twelfth century this was one of the most densely populated rural districts in England, seven miles from one of the wealthiest and most politically powerful monasteries in the country.

The two accounts agree on the core. Two children, brother and sister. Skin green over the whole body. Unknown speech. Unfamiliar clothing. Refusal of all food until raw broad beans were produced. The boy dies; the girl survives, loses the green colour as her diet normalises, and learns English.

They diverge everywhere else.

William dates the event to the reign of King Stephen, 1135 to 1154 — the civil war later called the Anarchy. Ralph gives no firm date. William says the children were found beside the wolf pits themselves, appearing suddenly after hearing a great sound like the bells of Bury St Edmunds Abbey while herding their father's cattle. Ralph says they followed cattle into a cave, wandered a long subterranean passage, and were guided out by the sound of bells. William says the girl afterwards married a man from King's Lynn, about forty miles away, and was still living there shortly before he wrote. Ralph says she served for years in the household of Sir Richard de Calne at Wykes, north of Woolpit, and was "very wanton and impudent."

Their homelands differ too. William's girl names the place St Martin's Land, a twilight country with a broad river and a luminous land visible beyond it. Ralph's girl describes a dim country in which everything, including the people, was green.

The sourcing differs as sharply. William says he had it from "unnamed persons," a number of trustworthy sources. Ralph names Richard de Calne, in whose house the children were said to have lived. Richard de Calne died in or before 1188, which means Ralph, writing in the 1220s, cannot have heard it from him directly.

Timeline

• 1086Domesday Book records Woolpit as a settlement of the Abbey of Bury St Edmunds.
• 1135-1154The reign of King Stephen, the period William assigns to the event.
• c. 1189: William of Newburgh, at Newburgh Priory in Yorkshire, writes the Historia rerum Anglicarum, including the green children.
• 1173The Battle of Fornham, near Bury St Edmunds. Royal forces destroy a rebel army containing large numbers of Flemish mercenaries. This date will matter to a theory proposed 825 years later.
• Before 1188Richard de Calne dies.
• 1220s: Ralph of Coggeshall writes the Chronicon Anglicanum, citing Richard de Calne as his source.
• 1586William Camden mentions the story in Britannia, following the first printed edition of Newburgh. He considers it a hoax.
• 1610A second edition of the Historia prints Newburgh's and Coggeshall's versions side by side for comparison.
• 1621Robert Burton, in The Anatomy of Melancholy, argues the story is true and that the children fell from the Moon.
• 1638Francis Godwin's The Man in the Moone is published posthumously, drawing on Newburgh; its lunar inhabitants venerate Saint Martin.
• 1850Thomas Keightley includes the tale in The Fairy Mythology, its first publication in English. From this point it is catalogued as a fairy legend.
• 1935Herbert Read, who had called the story "the norm to which all types of fantasy should conform," publishes the novel The Green Child.
• 1965John Macklin's Strange Destinies reports two green children arriving in the Spanish village of Banjos in 1887, complete with a mayor named Ricardo de Calno. No Spanish village called Banjos exists.
• September 1996The astronomer Duncan Lunan publishes "Children from the Sky" in Analog, proposing a matter-transmitter accident and a tidally locked planet whose only habitable zone is a narrow twilight band.
• 1998Paul Harris publishes the Flemish refugee hypothesis in Fortean Studies.
• 1999John Clark replies with "The Green Children: A Cautionary Tale," identifying problems in Harris's use of the sources.
• 2006Clark publishes two further papers, in Science Fiction Studies and Folklore, examining the tale's modern reception and the St Martin question.
• 2020Michał Madej publishes a comparative study of the two chronicles in Res Historica.
• April 2025Timothy R. Jones argues in a Medievalists.net feature that the details usually read as alien are evidence of an ordinary Northern European upbringing.

The Evidence

There is no material evidence: no contemporary administrative document recording the children's arrival, no name outside the chronicle tradition that can be shown to belong to someone who dealt with them, no burial. What exists is two texts and the question of whether they are independent.

That question matters more than any other, because it is the difference between one report and two. Medieval chroniclers copied each other freely and often without attribution. Madej argues they did not here: William wrote from Yorkshire, effectively the other side of England, decades before Ralph, and Ralph names his source while William does not. Clark has proposed a middle position — that Richard de Calne, or his household, may have been the ultimate source for both, which would make the accounts two branches from one root rather than two witnesses.

What the texts do contain, read closely, is a set of details that cut against the strangeness. The children knew they were brother and sister. They had been herding their father's cattle, which means pastoral farming existed where they came from and worked the same way it did in Suffolk. The girl learned English, understood what baptism was, and took a place in a household on the ordinary terms of service in exchange for keep. Timothy Jones's 2025 reading builds on exactly this: whatever else they were, these were children with a grounding in Northern European culture, not visitors from a world without agriculture or churches.

The green colour is the detail everyone remembers and the hardest to fix. A modern reader pictures a vivid pigment; Ralph's Latin says the whole surface of their skin was tinged green, which admits of a range. The candidate condition is hypochromic anaemia, historically called chlorosis or green sickness, which produces a greenish pallor and resolves with better food — and the colour did resolve as the children began eating normally. Against it, Madej makes a plain objection: much of the twelfth-century rural population was malnourished, and a tint common enough to be commonplace would not have been recorded as a marvel. Whatever the villagers saw was unusual to people who saw hungry children regularly.

The beans are firm as fact and slippery as meaning. Both chroniclers report them, and hungry children fixing on the one food they recognise needs no explanation at all. But Katharine Briggs, the major twentieth-century folklorist of English fairy tradition, called beans "the food of the dead," and that reading has travelled widely. Clark has pushed back, calling the identification unwarranted while allowing that beans are associated with the dead in many cultures; Madej notes they carried the opposite association too, rebirth and fertility.

The Theories

What Doesn't Add Up

The clothing. Both chroniclers report that the material was unknown to the villagers — and the villagers of Woolpit lived seven miles from Bury St Edmunds, in one of the busiest wool-producing regions in England, with continental trade and Flemish weavers in the district. These were not people unfamiliar with cloth. Either the detail is a literary flourish added by one or both writers, or it is the single hardest fact in the case for any refugee theory to absorb.

The language. Same objection, harder. Richard de Calne was a knight with a household; Bury St Edmunds was a monastic centre receiving pilgrims and traders. Flemish was not an exotic tongue in twelfth-century Suffolk. Whatever the children spoke, someone should have placed it.

The girl's later life is recorded two incompatible ways. William, writing earlier and further away, says she married and moved to King's Lynn and was alive as he wrote. Ralph, writing later and closer, and naming his source, says she spent years as a servant in de Calne's house and was badly behaved. Both cannot be right, and the one with the named source is the one written a generation after that source died.

And the whole case rests on a chain that cannot be tested. William reports "unnamed persons." Ralph reports a man who was dead before the earlier chronicle was written. Nancy Partner, a historian of twelfth-century historiography, has argued that hunting for natural explanations in episodes like this is useless to the study of William of Newburgh or of the Middle Ages generally — the question worth asking is why the story was told, not what happened. That is a defensible position. It is also an admission that the event itself is out of reach.

Latest Developments

Nothing new has been discovered. What has changed over the last two decades is the direction of the scholarship, which has moved away from solving the case and toward examining how it has been solved.

John Clark's work is the pivot. His 1999 "cautionary tale" dismantled the assumption that Harris's Flemish hypothesis had closed the matter; his 2006 papers traced how the story became science fiction and tested the St Martin associations that several folklorists had relied on, finding no evidence for Martin as a figure with Otherworld connections. His 2018 book remains the fullest modern treatment, and his consistent position is that the story may well be a historical report clothed in language that later readers found folkloric — and that forcing it into a single rational scheme requires brushing awkward details aside.

The most recent substantive contribution is Timothy R. Jones's April 2025 article, which inverts the usual method. Instead of asking what could make green children possible, it asks what the chronicles inadvertently record about ordinary life: that the children herded livestock, recognised kinship, understood baptism, and entered domestic service. On that reading the remarkable thing is not their strangeness but how much they had in common with the people who found them, and how comprehensively the retelling has erased it. Jones suggests they may have come from Northern Europe by trade ship and become disoriented — a variant of the refugee theory that does not depend on the Fornham date.

The case stays open because the evidence was always going to be insufficient, and everyone working on it now knows that. Two texts, no documents, no remains, and a chain of testimony with a dead man at the end of it. What can still be established is narrower: what the twelfth century thought it was recording, and how much of the mystery was manufactured by the seven centuries of retelling that followed.

Sources

1. Green children of Woolpit — Wikipedia, citing Clark 1999/2006, Harris 1998, Madej 2020, Cohen 2008, Briggs 1970
2. The Green Children of Woolpit — Examining the Medieval Records — Veriarch, 2025, updated 2026
3. Who Were the Green Children of Woolpit? A New Look at a Medieval Mystery — Timothy R. Jones, Medievalists.net, 2025
4. The Green Children of Woolpit: Going past the skin-deep explanations of a medieval legend — Medievalists.net, 2018
5. The Story About the Green Children of Woolpit According to the Medieval Chronicles of William of Newburgh and Ralph of Coggeshall — Michał Madej, Res Historica 49, 2020
6. The Green Children: A Cautionary Tale — John Clark, Fortean Studies 6, 1999
7. 'Agnes Barre': Green Child of Woolpit or historical fantasy? — John Clark
8. Placing the Green Children of Woolpit — academic paper via Academia.edu
9. Woolpit, Domesday Book entry — Open Domesday, 1086 survey
10. Bury St Edmunds Abbey: History — English Heritage
11. Battle of Fornham, 1173 — The Battlefields Trust
12. Children from the Sky — Duncan Lunan, expanding his 1996 Analog article
13. Green children of Woolpit (preprint) — WikiJournal Preprints, Wikiversity
14. The Man in the Moone — Wikipedia, on Francis Godwin's 1638 use of the story

GRB 221009A

On 9 October 2022 a burst of gamma rays saturated nearly every detector in orbit, and nobody agrees what shape it was.

The Hook

At 13:16:59.99 Universal Time on 9 October 2022, the Gamma-ray Burst Monitor aboard NASA's Fermi Space Telescope registered a trigger. Within seconds the count rate climbed past the point where the instrument could record individual photons. Detectors on Fermi, on Swift, on Konus-Wind, on the Chinese Insight-HXMT — instruments built in different decades, in different countries, to different designs — all ran into the same wall at roughly the same moment. The burst was brighter than the hardware built to measure bursts.

Fifty-three minutes later, Swift's Burst Alert Telescope reported a very bright transient at a position close to the plane of the Milky Way. Because of that position, the object was provisionally catalogued as a new Galactic X-ray source and given the designation Swift J1913.1+1946. It was not Galactic. It was a dying star roughly two billion light years away, and the two teams established within hours that they were looking at the same thing.

The event now carries an informal name that began as a joke and stuck: the BOAT, the Brightest Of All Time. Eric Burns and his co-authors argued in 2023 that, by geometric extrapolation of the fluence and peak-flux distributions of known bursts, an event this bright should occur roughly once every ten thousand years. Their phrasing was careful. It is almost certainly not the brightest burst in cosmic history. It may well be the brightest to have appeared in Earth's sky since human civilisation began.

What makes it a case file rather than a record entry is everything that has happened since. Four years of analysis have produced a supernova that is too faint, a jet whose width is disputed by a factor of three, a spectral line that no second instrument can confirm, and a claimed effect on Earth's upper atmosphere that a 2026 paper says was never there.

The Facts

The burst is formally GRB 221009A. Its redshift is z = 0.151, measured from spectroscopy of the afterglow — extraordinarily close for a long gamma-ray burst. Popular accounts give the distance variously as 1.9 or 2.4 billion light years; the difference is a choice of distance measure, not a disagreement about the object.

The energetics, as measured by the Konus-Wind and SRG/ART-XC instruments and published by Frederiks and colleagues in 2023: an isotropic-equivalent energy release of (1.2 ± 0.1) × 10^55 erg between 1 keV and 10 MeV in the rest frame, a total prompt fluence of 0.22 ± 0.02 erg per square centimetre, and a one-second isotropic luminosity of (2.1 ± 0.4) × 10^54 erg per second. Duration measurements vary with instrument and energy band: Konus-Wind gives a T90 of 284.0 ± 3.7 seconds, SRG/ART-XC 276.0 ± 5.4 seconds, Fermi-GBM roughly 300 seconds. Prompt emission of some kind continued for more than ten minutes.

The saturation problem is documented precisely. Fermi-GBM loses time-tagged event data above a total detector rate of 375,000 counts per second, and suffers pulse pile-up. The mission's official "bad time intervals" for this burst run from 219 to 277 seconds and from 508 to 514 seconds after trigger. The most intense part of the explosion was never directly measured by anything. Every figure above is a reconstruction.

On the ground, the Large High Altitude Air Shower Observatory in Sichuan was in a position to see the burst from the start. Its Water Cherenkov Detector Array recorded more than 64,000 photons above 0.2 TeV in the first 3,000 seconds — the first time the onset of a gamma-ray burst afterglow has been watched at these energies. Its KM2A array registered more than 140 photons above 3 TeV, with the highest at about 13 TeV.

The host galaxy is a low-metallicity dwarf, roughly 0.12 solar metallicity, with strong emission lines indicating recent star formation at the burst site.

Timeline

• 9 October 2022, 13:16:59.99 UTCFermi-GBM triggers. Detectors across the gamma-ray fleet begin to saturate.
• 13:20:50 UTCPhotons from the burst enter LHAASO's field of view; the TeV afterglow rises steeply, peaks around 245 seconds after trigger, and breaks steeply at roughly 670 seconds.
• 14:10:17 UTCSwift-BAT reports the source; it is catalogued as a possible Galactic transient, Swift J1913.1+1946. Within hours, ground observatories identify the optical afterglow and establish the redshift.
• October–November 2022Hayes and Gallagher report a sudden ionospheric disturbance in Earth's D-region, detected from Dublin. LHAASO reports photons up to 18 TeV in a rapid circular. Carpet-2 reports a photon-like air shower of 251 TeV, 4,536 seconds after the trigger.
• March 2023Burns and colleagues publish "GRB 221009A: The BOAT" and the once-in-10,000-years estimate.
• June 2023LHAASO publishes in Science: the TeV afterglow is fitted by a jet with a half-opening angle of about 0.8 degrees, interpreted as the core of a structured jet. In October its refereed analysis of the highest energies settles on 13 TeV, superseding the 18 TeV circular.
• November 2023Piersanti and the CSES-Limadou collaboration report, in Nature Communications, a perturbation of the ionospheric electric field at about 500 km altitude — claimed as the first GRB effect ever seen in the upper ionosphere.
• 2023–2024JWST spectroscopy detects the associated supernova, SN 2022xiw, a Type Ic-BL. It is unremarkable.
• 26 July 2024Ravasio and colleagues publish in Science a high-confidence emission line at about 12 MeV. In October the Fermi-GBM instrument team publishes an independent verification, names it the Ravasio Line, and states it cannot identify any instrumental effect that would produce it.
• 2025Sears and colleagues report a break in the light curve at 50 ± 10 days and a blue component in the late-time afterglow. Carpet-3 revises its 251 TeV event upward, to roughly 300 TeV.
• May 2026He and colleagues publish a reassessment concluding the upper-ionosphere signals were instrumental and geophysical artefacts. A separate re-analysis of the early spectra finds a jet break at 0.5–1.0 days, contradicting the 50-day figure.

The Evidence

The saturation record. Paradoxically, the strongest evidence for how bright the burst was is the record of instruments failing. The reconstruction back to a true brightness relies on those instrument models being correct.

The LHAASO photon catalogue. More than 64,000 photons above 0.2 TeV inside 3,000 seconds is not a marginal detection; it is the best-sampled TeV afterglow on record. The light curve is well resolved: a sharp rise at around 230 seconds, a peak at 245 seconds, a decay, and a steep break near 670 seconds. Whether that break is a jet break is precisely what is disputed.

The 12 MeV line. Ravasio and colleagues found a feature appearing roughly five minutes after trigger — after saturation had ended — persisting at least 40 seconds and peaking near 12 MeV, with a centroid that drifts downward to about 6 MeV by 340–360 seconds. The Fermi-GBM instrument team's independent check is unusually thorough: they exclude time-tagged-event saturation, pulse pile-up, scintillator afterglow, and every known fixed-energy background line, on the grounds that none of these can drift in energy. Their simulation-based significance for the 300–320 second interval exceeds 5.3 sigma, from twenty million simulations.

The supernova spectrum. JWST/NIRSpec observations at 168 and 170 rest-frame days, reported by Blanchard and colleagues, are well described by a supernova plus a power-law afterglow. The supernova component resembles previous GRB supernovae including SN 1998bw but is slightly fainter, with an inferred nickel mass of about 0.09 solar masses. There is no r-process signature.

The extinction problem. A 2026 re-analysis of VLT/X-Shooter and JWST spectra found a strong degeneracy between the assumed dust extinction along the line of sight and the inferred intrinsic spectral slope — a degeneracy that explains why published values had been irreconcilable. The best-fit extinction is A_V ≈ 4.4, with a plausible range of 3.8 to 5.05. Roughly four magnitudes of dust sit between this event and every instrument that measured it.

The Theories

What Doesn't Add Up

1. The jet width is unknown by a factor of three, and the jet break by four orders of magnitude. LHAASO's fit gives 0.8 degrees. Two-component models give a 0.6-degree Poynting-flux-dominated core plus a broader wing. Kusafuka and Asano give 1.7 degrees. As for when the jet break happened: LHAASO puts it at 670 seconds, the 2026 X-Shooter/JWST re-analysis at 0.5 to 1.0 days, Sears and colleagues at 50 ± 10 days, and Kusafuka and Asano say there was no jet break. These are not refinements of each other. They are incompatible.

2. The circumburst medium is required to be two different things. Radio and late-time modelling favour a wind profile, as a Wolf-Rayet progenitor should produce. The very-high-energy rise and several other analyses favour constant density. Kusafuka and Asano resolve this by inverting the expected profile. Meanwhile the 2026 explanation of the MeV line needs an extremely dense medium close in. The near-field and far-field requirements pull in opposite directions.

3. The supernova is too ordinary, or the dust is hiding it, and there is no way to tell. Photometric studies put SN 2022xiw at 10 to 70 per cent of SN 1998bw's luminosity. Blanchard and colleagues call it only slightly fainter, with 0.09 solar masses of nickel. A modelling paper arguing the supernova is obscured by the afterglow gets 0.23 solar masses of nickel. Sears and colleagues can only set an upper limit of 1.5 times SN 1998bw. With four to five magnitudes of line-of-sight extinction, itself degenerate with the spectral slope, whether the supernova is genuinely weak may not be answerable with the data that exist.

4. The electron spectral index comes out below 2. The 2026 re-analysis derives p = 1.89, which standard particle-acceleration theory does not comfortably produce. LHAASO's spectrum above 3 TeV is likewise harder than synchrotron self-Compton models predict. Two independent lines of evidence point at the same gap in the theory.

5. The most important spectral feature ever found in a gamma-ray burst was seen by one instrument, and always will be. Fermi's Large Area Telescope Low Energy was saturated; AGILE was not observing. GECAM-C extends the line to earlier intervals, but the Fermi-GBM team could not confirm those extensions as a line. There is no second independent measurement and there never can be — the photons are gone.

6. Nobody has revised the rate estimate. The once-in-10,000-years figure has stood unchallenged since 2023, while a separate 2022 estimate put an event this close and this energetic at "once in a century or longer." Both may be right; they compute different quantities. But the answer to "how unusual was this" ranges over two orders of magnitude depending on which question is asked.

Latest Developments

The most consequential recent result is negative. In May 2026, He, Li, Zhang, Forbes, Lei, Liu, Shi and Wang published a reassessment of the ionospheric claims, concluding that the reported upper-ionosphere responses are "fully attributable to other natural geophysical processes and instrumental artefacts." The CSES electric-field perturbation, they argue, is a photoemission artefact that recurs on every orbit and has nothing to do with the burst. This retracts, in effect, the single most widely reported "the burst affected Earth" claim. The lower-ionosphere VLF detection survives.

On the astrophysics, the picture has become less settled rather than more. The 2026 "Dusty Blue BOAT" analysis of VLT and JWST spectra found that the extinction-versus-spectral-index degeneracy explains the conflicting values published over the previous three years, and concluded that a wind medium is ruled out by the temporal evolution while a uniform medium works only if an early jet break at 0.5 to 1.0 days is invoked. That directly contradicts the 50-day break reported by Sears and colleagues in 2025, which in turn required a wind medium. The two most recent detailed analyses of the same object disagree on both the geometry and the environment.

LHAASO published a further result in 2025 on the spectral turnover in the afterglow, identifying two components in the first thirty minutes with the second peaking near 100 GeV. Carpet-3's revision of its high-energy event to roughly 300 TeV keeps the exotic-physics literature alive without resolving whether the event came from the burst at all. The Ravasio Line remains unexplained: pair annihilation, neutron capture and precursor-shell emission are all still on the table.

Four years after the brightest explosion in recorded history, the disagreements are sharper than they were in 2023.

Sources

1. GRB 221009A: The BOAT — Burns et al., ApJL 946 L31 / arXiv, 2023
2. Properties of the extremely energetic GRB 221009A from Konus-WIND and SRG/ART-XC observations — Frederiks et al., arXiv, 2023
3. A tera-electronvolt afterglow from a narrow jet in an extremely bright gamma-ray burst 221009A — LHAASO Collaboration, Science 380 1390 / arXiv, 2023
4. Very high-energy gamma-ray emission beyond 10 TeV from GRB 221009A — LHAASO Collaboration, Science Advances, 2023
5. Fermi Telescope finds new feature in brightest gamma-ray burst yet seen — Phys.org / NASA Goddard, 2024
6. NASA's Fermi Finds New Feature in Brightest Gamma-Ray Burst Yet Seen — NASA Science, 2024
7. Fermi-GBM Team analysis of the emission line in GRB 221009A — Burns, Lesage et al., arXiv, 2024
8. Evidence of an upper ionospheric electric field perturbation correlated with a gamma ray burst — Piersanti et al., Nature Communications 14 7013, 2023
9. Reassessment of Ionospheric Responses to GRB 221009A: Disentangling Instrumental, Illumination and Geophysical Effects — He et al., arXiv, 2026
10. A Significant Sudden Ionospheric Disturbance associated with Gamma-Ray Burst GRB 221009A — Hayes & Gallagher, arXiv, 2022
11. JWST Observations of the Extraordinary GRB 221009A Reveal an Ordinary Supernova — Blanchard et al., arXiv, 2023
12. JWST detection of a supernova associated with GRB 221009A without an r-process signature — Blanchard et al., Nature Astronomy, 2024
13. GRB 221009A/SN 2022xiw: A Supernova Obscured by a Gamma-Ray Burst Afterglow? — IOPscience, ApJ, 2024
14. Late-time HST and JWST Observations of GRB 221009A: Evidence for a Break in the Light Curve at 50 Days — Sears et al., ApJ 984 196 / arXiv, 2025
15. JWST's Dusty Blue BOAT — GRB 221009A — Khang, Lamb et al., arXiv, 2026
16. TeV Afterglow of BOAT GRB without Jet Break — Kusafuka & Asano, MNRAS / arXiv, 2025
17. LST-1 observations of GRB 221009A: insights into its late-time VHE afterglow — CTAO/LST Collaboration, arXiv, 2025
18. Carpet-3 detection of a photon-like air shower above 100 TeV in coincidence with GRB 221009A — Dzhappuev et al., Phys. Rev. D 111 102005 / arXiv, 2025
19. Featured Image: Hubble and JWST Check Up on the BOAT Gamma-Ray Burst — AAS Nova, 2025
20. Observation of the spectral turnover in the afterglow emission of GRB 221009A — LHAASO Collaboration, Astronomy & Astrophysics, 2025

The Jamison Family

A locked truck on a well pad, about $32,000 inside, a starving dog, and three people who walked into the Sans Bois and did not come back.

The Hook

In mid-October 2009, hunters came across a white Ford F-150 parked on a remote well pad in the Sans Bois Mountains of Latimer County, Oklahoma. It had been there for something like eight or nine days. Inside were two wallets, a purse, identification cards, cell phones, a GPS unit, the family's jackets, and a quantity of cash that has been reported ever since as roughly thirty-two thousand dollars. Also inside was a small dog named Maisie, still alive and severely emaciated.

The people who owned all of it — Bobby Dale Jamison, 44, his wife Sherilynn, 40, and their six-year-old daughter Madyson — were not there. They had driven up from Eufaula, about thirty miles away, to look at a forty-acre plot of land they were considering buying.

More than three hundred volunteers and dozens of officers searched the surrounding mountains. Aircraft and dogs were used. Nothing was found. Four years later, in November 2013, two deer hunters found skeletal remains 2.7 aerial miles northwest of where the truck had been parked. It took the Oklahoma State Medical Examiner until July 2014 to confirm the identifications, and when the office did, it stated that both the cause and the manner of death were unknown, because the remains were incomplete.

That is where the case still sits. Three people died within walking distance of their own vehicle, in terrain that had been searched intensively, and there is no forensic finding as to how.

The Facts

Bobby Dale Jamison and Sherilynn Leighann Jamison lived in a house on Eufaula Lake in Eufaula, Oklahoma, with their daughter Madyson Stormy Star Jamison. Then-Latimer County Sheriff Israel Beauchamp told the Associated Press that both adults were disabled and did not work. Reporting at the time noted the family had originally come from the Oklahoma City area and travelled frequently to timeshare properties.

The land they went to see was a forty-acre plot for sale in the northern part of Latimer County, near Red Oak, in the Sans Bois range. Accounts of where exactly the truck was found vary. Contemporaneous Associated Press reporting places it at a remote well pad site near the plot the family was considering; other accounts place it on a dirt road nearer Kinta. The AP account, being closest in time to the event and sourced to the sheriff, is the more reliable.

The date they were last seen is 8 October 2009, based on Sheriff Beauchamp's statement to the AP that a landowner in the Red Oak area saw them that day. One 2014 broadcast account gives 9 October; this appears to be an error. The truck was found on 16 or 17 October — reporting differs — by hunters, and the large-scale search was launched on 17 October.

The contents of the truck are the most-repeated element of the case and the least officially documented. Wallets, purse, phones, identification and the dog are consistent across every account. The cash figure of "about $32,000" appears in essentially all coverage but has never, in any material reviewed for this file, been attributed to a named official or agency document; one recent secondary account gives a precise $32,690. Whether the keys were in the vehicle is genuinely disputed — one account says they were missing, another that they were under the driver's seat, and neither is sourced to an investigator. The dog's breed is likewise unverified; the AP described only "Madyson's small dog."

Two items that were expected and never recovered: a brown briefcase, seen on the family's own home security footage being loaded into the truck, and Sherilynn's small-calibre handgun. Beauchamp told The Oklahoman of the briefcase, "I think it might have held a lot of cash."

Timeline

• 7 or 8 October 2009The family's home security system in Eufaula records Bobby and Sherilynn making repeated trips between house and truck, loading it. Sources differ on the exact date of the recording.
• 8 October 2009A landowner near Red Oak sees the family, according to Sheriff Beauchamp. This is the last confirmed sighting.
• 16 or 17 October 2009Hunters find the F-150 on a well pad. Maisie is alive inside, starving.
• 17 October 2009A search involving more than 300 volunteers and dozens of law-enforcement officers begins. Heavy rain has fallen in the area. Nothing is found.
• December 2009Bob Dean Jamison, Bobby's father, dies, roughly two months after the disappearance.
• February 2013KOKH Fox 25 broadcasts an account of an anonymous caller to a friend of Sherilynn's, and the family's security footage becomes widely known.
• 16 November 2013Two deer hunters find skeletal remains of two adults and one child, 2.7 aerial miles northwest of the truck site. The Latimer County and Haskell County Sheriff's Offices, the OSBI, the FBI and the State Medical Examiner respond.
• Late November 2013The Oklahoman reports the existence of a 2009 protective-order petition filed by Bobby against his father.
• 3 July 2014The Oklahoma State Medical Examiner positively identifies the remains as the Jamison family. Cause of death: unknown. Manner of death: unknown.
• 2013 to the presentThe case remains open and unsolved. No suspect has ever been publicly named.

The Evidence

The vehicle. Everything the family would plausibly need to leave, travel or survive was inside it: money, phones, identification, outerwear, navigation. Beauchamp's own reading of the scene, given to The Oklahoman, was based on how the truck was parked and what was in it — he came to believe they were stopped and got out to meet someone they recognised. The AP reported at the time that there were no signs of foul play and that it appeared the family had intended to return to the truck.

The security footage. Recorded by the family's own system at the Eufaula house. It shows the two adults making a large number of trips — one account puts it near twenty — between house and vehicle, in silence, packing methodically and slowly. The description repeated by investigators and reporters is of a daze or trance. KOKH Fox 25 reported in 2013 that investigators said the movements "appear to show the family in some sort of daze, which could point to drug use." Both adults appear unusually thin. A friend of Sherilynn's, Niki Shenold, has stated she believes a third person appears in the footage; this is her reading, not an investigative finding.

The photograph of Madyson. A photograph of the girl was recovered from a phone found in the truck, described as taken on the mountain. Oklahoma Cold Cases records the investigatively relevant detail plainly: Madyson is wearing a red hooded jacket that she did not have with her at the time of the disappearance. The widely circulated characterisation — that she is looking away from the camera, arms crossed, unhappy — comes from friends and family, not from investigators. No source reviewed for this file gives a verified timestamp for the image, and the online claim that it was taken after her parents were dead has no evidentiary basis whatsoever.

The letter. Investigators found an eleven-page letter in the truck, written by Sherilynn to Bobby, listing grievances against him and, in one reported phrase, calling him a hermit. It is a document about a marriage under strain. It is not, despite frequent conflation, a document about spirits.

The remains. Two adults and one child, skeletal and incomplete, found by hunters 2.7 aerial miles northwest of the truck. Several accounts state the three were lying side by side and face down; this detail is carried by secondary sources rather than by the OSBI statement or the AP wire copy. The medical examiner's finding, issued in July 2014, was that the cause of death could not be determined because of the incompleteness of the remains, and that the manner of death was likewise unknown.

The Theories

What Doesn't Add Up

1. The search failed, twice over. More than three hundred people, with dogs and aircraft, searched terrain in which the bodies lay 2.7 aerial miles from the starting point. They were not found for four years, and then only by hunters. Either the remains were moved there after the search, or the search missed them at close range.

2. Nothing was taken. Whatever happened, roughly thirty-two thousand dollars in cash stayed in the vehicle. That rules out the simplest motive for a roadside encounter in a remote place, and leaves either a motive unrelated to money or no third party at all.

3. The briefcase and the handgun are gone. Both were known to be with the family. Neither has ever been recovered. They are the only two things unaccounted for, which makes them either the most important objects in the case or a bookkeeping artefact of an unrecorded packing decision.

4. The cash figure has no official source. The number that defines this case in public memory — $32,000 — cannot be traced, in the material reviewed here, to any agency statement or court document. Whether the keys were in the truck is similarly untraceable. A case built largely on the contents of one vehicle rests on an inventory nobody official has published.

5. The medical examiner could not determine anything. No cause, no manner, no toxicology. The remains were incomplete and had been exposed for four years. This means every theory above, including the most ordinary one, is unfalsifiable on the physical evidence. That is unusual for a case that attracts this much attention.

6. The most quoted detail about the family — the trance-like video — is an interpretation. Investigators offered it as consistent with drug use. It is equally consistent with two exhausted, chronically unwell adults packing a truck at the end of a long day, recorded by a low-frame-rate security camera. No one has produced anything that distinguishes between those readings.

Latest Developments

There is very little to report, which is itself the current state of the case.

The Latimer County Sheriff's Office stated in 2013 that the case was open and active but declined to comment on any theory. Oklahoma Cold Cases, a state non-profit, still lists the case as unsolved and gives the Latimer County Sheriff's Office as the investigating agency. A 2025 regional account states that the Oklahoma State Bureau of Investigation continues to oversee the case, which remains open but inactive for lack of new leads; this is a media characterisation rather than an agency statement. The Jamison case does not appear on the OSBI's public cold-case index.

No new search, excavation, forensic re-examination, reward or agency press release dated between 2020 and 2026 could be identified for this file. The case has been covered by the Investigation Discovery series Disappeared, by BuzzFeed Unsolved, by Crime Junkie, and by The Path Went Chilly in a two-part treatment in September 2024. None of it produced new investigative material; all of it restates the 2009–2014 record. The only genuinely new information generated by media in the case's history remains the 2013 KOKH broadcast.

A note on one strand of the case, because it is frequently reported without its qualifiers. In 2009, roughly six months before the family vanished, Bobby Jamison filed a petition for a protective order in McIntosh County against his own father, Bob Dean Jamison, alleging threats to kill him and his family on two occasions and, separately, alleging involvement in criminal activity. Those are allegations made by one man in a civil filing. A judge heard testimony and dismissed the petition on 18 May 2009. Bob Dean Jamison was never charged with anything, was never publicly named a suspect, died in December 2009, and — as The Oklahoman reported — police do not believe he was involved in the disappearance. A separate property lawsuit between Bobby and his parents was still ongoing when the family vanished. The conflict is a documented fact about the family. It is not a finding about the deaths.

Sources

1. Jamison family deaths — Wikipedia, accessed 2026
2. ME: Remains Found In Latimer County Identified As Family Missing Since 2009 — News9 (KWTV), 2014
3. Found remains may belong to missing Oklahoma family — Associated Press via Fox News, 2013
4. Found remains could belong to missing Okla. family — Associated Press via Enid News & Eagle, 2013
5. Missing Oklahoma Family Threatened Before Vanishing — Associated Press via Public Radio Tulsa, 2013
6. Oklahoma family was threatened before vanishing, records say — Associated Press via Tulsa World, 2013
7. Skeletal remains found by hunters in Okla. could belong to Jamison family — CBS News, 2013
8. Missing Family Mystery Expands Thanks to Video and Anonymous Tipster — KOKH Fox 25 Oklahoma City, 2013
9. Missing Oklahoma Family, Demons and Drugs — Guardian Liberty Voice, 2013
10. Bobby Jamison — Oklahoma Cold Cases Inc., 2024
11. Madyson Jamison — Oklahoma Cold Cases Inc., 2024
12. Sherilynn Leighann Jamison — The Charley Project, undated
13. Cold Case Files — Oklahoma State Bureau of Investigation, 2026
14. Cold Case Files: Discovery of bodies near Kinta in 2009 leaves more questions than answers — Today in Fort Smith, 2025
15. The Chilling Story Of The Jamison Family — All That's Interesting, 2026

USS Cyclops

A 542-foot Navy collier left Barbados on 4 March 1918 with one working engine and about 10,800 tons of ore, and was never seen again.

The Hook

On 4 March 1918, the USS Cyclops cleared Barbados for Baltimore. She was a fleet collier, 542 feet long, carrying something in the region of 10,800 long tons of manganese ore. Her starboard engine had a cracked cylinder and did not work. She was riding below her Plimsoll line — the load mark painted on a ship's side to show when she is too deep — which is why she had put into Barbados in the first place, unscheduled. While there she took on more coal and provisions, weighing her down further.

She was due at Baltimore on 13 March. She never arrived. No distress signal was ever received. No wreckage, no lifeboat, no body, and no object identifiable as hers has ever been recovered. Roughly 306 people were aboard.

The Secretary of the Navy, Josephus Daniels, wrote in his 1918 annual report that there had been no more baffling mystery in the annals of the Navy. That framing has survived a century largely because the alternatives keep failing. German records were checked after the Armistice and showed nothing. The weather data that would put her in a storm depends on a sighting the sighting ship's own master denied. And in 1941, two of her three sister ships vanished in the same way, without trace, carrying the same kind of dense ore — which is why this case file is not really about a triangle.

The Facts

Cyclops was built by William Cramp & Sons of Philadelphia — keel laid 2 June 1909, launched 7 May 1910, placed in service 7 November 1910 with a civilian crew under a licensed master as part of the Naval Auxiliary Service. She was commissioned into the Navy proper on 1 May 1917. She displaced 19,360 tons fully loaded, was 65 feet in the beam, and made a shade under 15 knots on trial with twin screws and vertical triple-expansion engines.

She is commonly described as a Proteus-class collier, though the most detailed technical record available classes her as her own type. The point matters only because the design, whatever it is called, will come back. Her capacity is itself disputed: one construction record gives 11,600 tons of coal across thirteen holds, while the Navy's own ship history gives a maximum of 8,000 long tons, which would make her badly overloaded in March 1918. No source reconciles the two.

She was assigned to the Naval Overseas Transportation Service on 9 January 1918 and sent to Brazilian waters to fuel British warships in the South Atlantic. She left Rio de Janeiro on 15 or 16 February, called at Bahia on 20 February, and departed for Baltimore on 21 or 22 February with no further stops planned. On 3 March she put into Barbados anyway.

The number of people aboard is the Navy's own soft figure. The Naval History and Heritage Command gives 306 — 250 crew and 56 passengers — and explicitly notes that the numbers vary between sources. A construction record gives 309. Contemporaneous press in April 1918 gives 293. Among the passengers was Alfred Louis Moreau Gottschalk, the US Consul-General at Rio de Janeiro, travelling home to enlist in the Army; five court-martial prisoners were also aboard, one serving a life sentence for complicity in a murder.

One correction, because it appears in nearly every account of this case: Cyclops was not the largest non-combat loss of life in US Navy history. The Naval History and Heritage Command states, twice, that the distinction belongs to the frigate USS Insurgent, lost without trace in 1800 with about 340 aboard, and that the worst non-combat loss overall was Typhoon Cobra in December 1944, which sank three destroyers and killed 790 men.

Timeline

• 15 or 16 February 1918Departs Rio for Bahia, five court-martial prisoners aboard. On this leg the starboard high-pressure cylinder head blows off; she continues on one engine.
• 20 February 1918Arrives Bahia, overrunning the port by 48 miles against the navigator's protest. Departs for Baltimore on the 21st or 22nd, no stops scheduled.
• 3 March 1918Puts into Barbados unscheduled, riding below her Plimsoll line. The US Consul, Charles Ludlow Livingston, records that the master says the engines are in very poor condition, has insufficient funds, and is "unusually reticent."
• 4 March 1918Departs Barbados. This is the last confirmed position.
• 9–10 March 1918The tanker Amolco reports a heavy gale off Virginia. A rumoured sighting of Cyclops by Amolco is denied by Amolco's own master.
• 13 March 1918Due at Baltimore. Does not arrive.
• 1 June 1918Assistant Secretary of the Navy Franklin D. Roosevelt declares Cyclops officially lost and all hands deceased. The recorded date of death for all aboard is 14 June 1918.
• November 1918Days after the Armistice, Chief of Naval Operations Admiral William S. Benson cables Admiral William S. Sims in Europe, directing that every effort be made to obtain information from German sources. Nothing is found.
• January 1920Lieutenant Commander Mahlon S. Tisdale publishes "Did the Cyclops Turn Turtle?" The Navy Department adopts a capsize theory.
• November and December 1941USS Proteus and USS Nereus, sold into Canadian commercial service, vanish without trace in the same region, both carrying bauxite ore.
• 1968–1969A wreck located 70 miles off Cape Henry is dived by Master Diver Dean D. Hawes, who finds an old cargo ship apparently broken in two. Conrad A. Nervig, a watch officer aboard Cyclops on her final outbound leg, publishes a first-hand account concluding she broke in two.
• 1974 and c.1983Navy divers and, separately, a NUMA expedition search the Virginia Capes wreck sites. Both come up negative.

The Evidence

Livingston's telegram. The US Consul at Barbados wired the Secretary of State on 2 April 1918. It is the last documentary record of the ship and her people, and it survives in full. Livingston records that the master needed 600 tons of coal, had insufficient funds and asked the consul to pay; that the engines were in very poor condition; that he was "unusually reticent"; that there were 231 crew exclusive of officers and passengers; and that Consul-General Gottschalk was aboard. He also noted that many Germanic names appeared, then closed with a sentence that has done more damage to this case than any other: "While not having any definite grounds I fear fate worse than sinking though possibly based on instinctive dislike felt towards master."

Nervig's eyewitness account of the hull working. This is the strongest physical evidence in the case, and it comes from an officer who stood watch aboard her weeks before she disappeared. On a darkened deck he heard a sound like metal plates rubbing together. Investigating, he found that where steam and water pipes passed through web irons or touched the hull, the movement was visibly perceptible. In daylight, sighting fore and aft along the well deck, he could see the deck amidships rising and falling as the ship conformed to the seas. He reported it to Worley, whose answer was: "Son, she'll last as long as we do."

The engine survey. Worley reported the cracked starboard cylinder before sailing. A survey board confirmed it — and recommended the ship return to the United States anyway. She ran on one engine, her speed of advance down to roughly 240 nautical miles a day.

The 1932 inspection of USS Jason. On a sister ship, a seaman chipping rust knocked his hammer clean through her side. Inspection found that the heavy longitudinal I-beams running inside the skin — the ones buried in the coal bunkers, where they could not be seen — had had their three-foot webs and inner flanges entirely destroyed by corrosion, leaving only weakened plating tying the bow to the rest of the ship. Jason was thereafter routed light and kept south of all storm tracks. This is documented structural failure in the same design, fourteen years later.

The two sister ships. Proteus and Nereus were both sold into Canadian commercial service in 1941 and both vanished without trace, with full crews and cargoes of bauxite ore, within about three weeks of each other. Neither loss was the result of German submarines.

The Theories

What Doesn't Add Up

1. Three ships of one design vanished without trace, and only one of them is treated as a mystery. Cyclops in 1918, Proteus and Nereus in 1941 — all carrying dense ore, none lost to enemy action, none ever found. A theory that explains only the first is the wrong shape.

2. Nothing floated. Not one item identified as hers has ever been recovered, in a century, from a ship carrying 306 people with boats in their davits. Both surviving mechanisms — a hull break with both sections going vertical, or a capsize in seconds with everything secured for sea — predict exactly this. Almost nothing else does.

3. The two eyewitness officers contradict each other on the central question. Tisdale, who served in the class and rode this ship, says the cargo shifted. Nervig, who was aboard on her final outbound leg, says manganese ore would not have shifted at all. Both arguments are first-hand and both are technically literate.

4. Several details that circulate as fact do not survive checking. The sentence quoted in nearly every account — that the disappearance "has been one of the most baffling mysteries in the annals of the Navy" — is reproduced identically across dozens of outlets, always attributed to "the official Navy record" and never to a document or archive reference; the verifiable official statement is Daniels' in the 1918 annual report, which is similar but not identical. No contemporary document describes Cyclops as "down by the head"; the evidence is deep loading, not bow-heavy trim. And "largest non-combat loss of life in US Navy history" is corrected twice by the Navy's own historians.

5. Her number of dead is unknown. 293, 306 and 309 are all in circulation, from the press, from the Navy, and from the construction record respectively. A century on, nobody can say with certainty how many people were aboard.

Latest Developments

There has been no new evidence, no declassification and no expedition in the last five years.

The most recent official treatment is H-060-2, "Lost With All Hands," published by the Naval History and Heritage Command in April 2021. It restates 306 aboard with the caveat that sources vary, and notes that although the loss occurred in wartime there is no evidence enemy action had anything to do with it. In October 2021 the National Archives published a piece noting that it holds her Booklet of General Plans — effectively the only tangible thing left of the vessel.

In November 2024, the Doughboy Foundation published an appeal by Marvin W. Barrash, great-nephew of one of her firemen and author of a 703-page history of the ship, arguing that the word "mystery" should be retired, that any search must be for study rather than salvage because the site is the grave of over three hundred Americans, and calling on those with the means to mount one. Nothing traceable came of it.

The search history is entirely negative. In 1968 a wreck was located 70 miles off Cape Henry; Master Diver Dean Hawes dived it and found a very old cargo ship apparently broken in two. Reading Nervig's article the following year, he was struck by the resemblance, and wrote carefully that he could not say the ship was Cyclops but that the possibility existed. Navy divers examined a candidate wreck in 1974 and, after recovering the ship's bell, determined it was not her. Around 1983 Clive Cussler funded a NUMA expedition working from the 1968 dive coordinates; it found nothing. Cussler afterwards wondered whether the wreck Hawes had dived was Proteus or Nereus.

Barrash's own view is that she lies in the Puerto Rico Trench, more than 27,000 feet down — which, if correct, places her beyond any search yet contemplated. Claims of discovery in 2017 and 2024 have been examined and rejected; in one case the recovered artifacts proved to be of English manufacture.

The National Museum of the US Navy states the position without qualification: her wreck has never been found, and the cause of her loss remains unknown.

Sources

1. H-016-4: World War I — USS Cyclops and Her Sisters — Naval History and Heritage Command, 2018
2. H-060-2: Lost With All Hands — Naval History and Heritage Command, 2021
3. H-019-5: "Black Sunday" and the Battle of Orleans — Naval History and Heritage Command, 2018
4. USS Cyclops Mystery — National Museum of the U.S. Navy, undated
5. USS Cyclops (1910-1918) — National Museum of the U.S. Navy, undated
6. "…her loss remains unknown…": The Mysterious Disappearance of the USS Cyclops (AC-4) — National Archives, The Unwritten Record, 2021
7. The Cyclops Mystery — U.S. Naval Institute Proceedings (Conrad A. Nervig), 1969
8. Did the Cyclops Turn Turtle? — U.S. Naval Institute Proceedings (Mahlon S. Tisdale), 1920
9. Comment and Discussion, January 1970 — U.S. Naval Institute Proceedings (Potter, Hawes, van Deurs), 1970
10. The Mysterious Fate of the Cyclops — U.S. Naval Institute Naval History, 2020
11. USS Cyclops — The Deadliest Unsolved Mystery in the Navy — U.S. Naval Institute Naval History, 2021
12. The Unanswered Loss of USS Cyclops – March 1918 — Naval Historical Foundation (Capt. Lawrence B. Brennan), 2013
13. USS Cyclops (Fleet Collier No. 4) — shipscribe.com (Stephen S. Roberts), 2012
14. USS Cyclops (II) (Fuel Ship #4) — NavSource Online, undated
15. William Robert Wolf and USS Cyclops — The Mariners' Museum and Park, 2021
16. Cyclops expedition — National Underwater and Marine Agency (Clive Cussler), 1983
17. Marylander Tries to Preserve Memory of USS Cyclops — NUMA (Ellsworth Boyd), 2018
18. U.S.S. Cyclops – It's Time for Answers — Doughboy Foundation (Marvin W. Barrash), 2024
19. Bermuda Triangle Mystery: What Happened to the USS Cyclops? — HISTORY, 2018
20. USS Cyclops — Wikipedia, accessed 2026

The Beaumont Children

Three siblings caught a bus to a crowded beach on Australia Day 1966 and were seen playing with a stranger. Nothing of them has ever been found.

The Hook

The distance from the Beaumont house at 109 Harding Street, Somerton Park, to the beach at Glenelg is about three kilometres. On a hot Australia Day morning in 1966 that was a five-minute bus ride, and it was a journey the three Beaumont children had made before. Jane was nine, Arnna was seven, Grant was four. Their mother gave them six shillings and sixpence — enough for the fares and for lunch — and expected them back on the noon bus.

They did not come back on the noon bus, or the two o'clock. When their father returned early from a sales trip at around three, he drove straight to the beach and could not find them. At about half past five, or by one contemporaneous account 7:20 in the evening, the parents walked into Glenelg police station.

What followed was, at the time, the largest search in South Australian history. Police and volunteers went through beaches, dunes, stormwater drains and suburbs. The Patawalonga boat haven was drained. Airports, rail lines and interstate roads were watched. None of it produced anything. Sixty years on, not one of the seventeen items the children had with them — clothing, towels, a bag — has ever been recovered, and neither have the children.

What makes this case different from an ordinary disappearance is that there are witnesses, and the witnesses agree. Several people saw the three children that morning playing with an adult man. They described him consistently. He has never been identified.

The Facts

Jane Nartare Beaumont was born on 10 September 1956, Arnna Kathleen on 11 November 1958, and Grant Ellis on 12 July 1961. They lived with their parents, Grant "Jim" Beaumont, a former serviceman and taxi driver working as a travelling salesman, and Nancy Beaumont, née Ellis.

On 25 January 1966, in the middle of a heatwave, Jim dropped the children at Glenelg Beach before leaving on a three-day sales trip. The next morning they asked to go again. Accounts of the departure time differ: the formal police record gives around 10 a.m., and their mother at one point said they left on an earlier bus, at 8:45. That discrepancy is not a small one — it is roughly eighty minutes of a four-hour window — and it has never been resolved.

Several witnesses at Colley Reserve, the grassed area beside the beach, saw the three children playing in the sprinklers with a man. The description is unusually consistent: tall, thin-faced, sun-tanned, medium to athletic build, fair to light-brown hair, in his mid-thirties, wearing swim trunks. The children appeared relaxed with him. He approached at least one witness to ask whether anyone had been near the children's belongings, saying their money was missing. He then went to change while the children waited for him. Police estimated the group left the reserve together at around 12:15.

At Wenzel's cake shop on Moseley Street, a shopkeeper who knew the children well served Jane, who paid with a one-pound note. Two things about that transaction mattered to investigators. Nancy had given the children six shillings and sixpence, not a pound. And the children bought a meat pie, which the shopkeeper said they had never bought before.

A postman who knew the children gave a statement that he had seen them walking alone along Jetty Road, hand in hand and laughing, at about 2:55 p.m. Two days later he contacted police to say he now believed the sighting had been in the morning rather than the afternoon. A separate witness, Daphne Gregory, reported seeing the children with the man at about 2:45 p.m., and said he was carrying an airline bag resembling one of Jane's.

Nancy later remembered a remark Arnna had made at home before the disappearance: that Jane had "got a boyfriend down the beach". At the time she took it to mean a playmate.

Timeline

• 25 January 1966Jim Beaumont drops the children at Glenelg Beach, then leaves on a sales trip to Snowtown.
• 26 January 1966, morningThe children take the bus to Glenelg. The departure time is recorded as about 10 a.m. in the police file; Nancy Beaumont at one stage said 8:45 a.m.
• Late morningMultiple witnesses see the three children at Colley Reserve with an unidentified man in swim trunks.
• About 12:15 p.m.Police estimate the children and the man leave the reserve together.
• Around middayJane buys pasties and a meat pie at Wenzel's with a one-pound note she should not have had.
• About 2:45 p.m.Daphne Gregory reports seeing the children with the man, who is carrying an airline bag.
• About 2:55 p.m.A postman reports seeing the children walking alone on Jetty Road. He revises the timing two days later.
• Afternoon and eveningJim returns home early, searches the beach, then both parents search streets and friends' houses. The disappearance is reported at Glenelg police station.
• 29 January 1966The Patawalonga boat haven is drained after a reported sighting near it. Nothing is found. The Adelaide Sunday Mail runs the headline "Sex crime now feared". The initial reward is A£250.
• 8 November 1966The Dutch clairvoyant Gerard Croiset arrives in Australia. He identifies a site at a Somerton Park warehouse; public subscription raises the money to demolish and excavate it. Nothing is found. The site is searched again in 1996, with the same result.
• 1968The parents receive letters purportedly from Jane and from a man claiming to be the children's guardian, postmarked Dandenong, Victoria. A meeting is arranged and nobody appears.
• 1992Improved fingerprint techniques identify the letters as a hoax written by a teenager, by then a 41-year-old man. He is not charged, because of the time elapsed.
• 2013The book The Satin Man prompts scrutiny of the late Adelaide businessman Harry Phipps. A radar survey and a one-square-metre excavation at his former New Castalloy factory in North Plympton find nothing.
• 2 February 2018A second, nine-hour excavation at the same site, following a Major Crime review and a Channel Seven investigation, finds animal bones and rubbish. Detective Superintendent Des Bray: "There is nothing that has been located today that is in any way connected with the disappearance of the Beaumont children."
• 16 September 2019Nancy Beaumont dies in an Adelaide nursing home, aged 92.
• 9 April 2023Jim Beaumont dies in Adelaide, aged 97.
• 22 February – 1 March 2025A third excavation at the Castalloy site, privately organised by the MP Frank Pangallo, moves roughly 3,000 cubic metres of soil to depths of four metres. Nothing connected to the case is found.
• 26 January 2026Sixty years. Pangallo throws three yellow roses into the water at Glenelg and writes to the City of Holdfast Bay asking for a permanent memorial.

The Evidence

There is almost none. That is the defining feature of the case.

The witness descriptions. The single strongest piece of material in the file is that several independent people at Colley Reserve described the same man in the same way. It is consistent enough that South Australia Police have worked from it for six decades, and it is the yardstick against which every subsequent suspect has been measured. What it cannot do is identify anyone.

The one-pound note. The children left home with six shillings and sixpence and Jane spent a pound. The shopkeeper knew them and knew their usual purchases. Police concluded the money came from someone else — the clearest indication in the case that an adult was not merely nearby but actively providing for the children.

The behaviour. The parents described Jane in particular as shy. Investigators took the children's evident comfort with the man to mean they had met him before, possibly on an earlier visit to the beach, and Arnna's remark about Jane's "boyfriend down the beach" fits that reading. This is inference, not proof, but it is the inference the investigation has rested on.

Nothing physical. No bodies. No clothing, towels or bag, despite seventeen listed items. No scene, no vehicle, no forensic trace of any kind. Three separate excavations at the site most often named as a burial place have produced animal bone and factory rubbish.

A contaminated fragment. In March 2025, volunteers working with the journalist Bryan Littlely recovered a juvenile-sized pelvic bone fragment with green glass embedded in it from a property at Yatina once owned by a convicted paedophile. The forensic archaeologist Maciej Henneberg assessed it as most likely from a child. South Australia Police have stated that the handling and cutting of the item before it was handed over "has contaminated and compromised the forensic process", and that it is not known when testing will be completed. Early indications reported in 2026 point towards the 1973 Adelaide Oval abductions rather than this case.

The Theories

What Doesn't Add Up

1. The postman changed his time. His original statement placed the children walking alone on Jetty Road at about 2:55 p.m. — already an hour late, apparently unhurried, and without the man. Two days later he said it had been the morning. Both versions cannot be right, and the case looks materially different depending on which is.

2. The departure time is disputed by the family's own account. About eighty minutes of the children's last day is unaccounted for depending on whether they left at 8:45 or 10 a.m. That gap has sat unresolved in the file for sixty years.

3. The man wanted to be seen. He played with the children in the open, in daylight, at a crowded reserve on a public holiday, and voluntarily spoke to a bystander about missing money — drawing attention to himself. That is not the behaviour of someone planning what police believe happened next, which is why investigators concluded the children already knew him.

4. Seventeen items and not one of them has surfaced. Towels, clothing, a bag. Objects of that kind usually turn up. Nothing has, anywhere, in six decades.