Orbs & Ghost Lights

The least mysterious mystery

Ghost lights are the best-attested, least mysterious mystery in the whole file, and the most instructive. Every generation reports the same thing — a bright point of light near the horizon that seems to move, split, and follow you — attaches it to whatever technology or theology is current, and eventually builds a viewing platform. When somebody points an instrument at it, the answer is almost always distant artificial light bent by a temperature inversion. That resolution is not a letdown. It is the only explanation that accounts for why sincere people are not lying.

The load-bearing document for the entire territory is a federal pamphlet. George Rogers Mansfield of the U.S. Geological Survey spent two weeks near Brown Mountain, North Carolina in March and April 1922, observing on seven evenings, four of them past midnight, with a fifteen-inch plane table and a telescopic alidade. He occupied three stations — near Loven's Hotel, on the east slope of Gingercake Mountain, and at Blowing Rock — and spent two nights on the summit of Brown Mountain itself, where rain and fog defeated him and he saw nothing at all. His conclusion was released to the press in 1922 as Press Notice 14328 and reprinted verbatim in 1971 as Circular 646:

"In summary it may be said that the Brown Mountain lights are clearly not of unusual nature or origin. About 47 percent of the lights that the writer was able to study instrumentally were due to automobile headlights, 33 percent to locomotive headlights, 10 percent to stationary lights, and 10 percent to brush fires."

Note the qualifier he built into his own sentence. Those percentages cover only the lights he fixed by instrument; he thought the total seen may have been nearly twice the number recorded. A century of retellings has dropped that clause. Mansfield did not.

The finding that matters sits a page or two earlier, and almost nobody quotes it. Of the twenty-three lights he plotted, seven proved to be locomotive headlights and an eighth probably did.

"None of these seven lights cast a beam or possessed any special quality that distinguished it from the other lights observed, and only one of them (No. 16) was known at the time of observation to be a locomotive headlight. The other six were identified as such only after the lines and angles of the record had been plotted, profiles had been drawn, and train schedules had been checked some days after the observations were made."

His party included "persons long familiar with the light." One in seven. That is the entire epistemology of eyewitness testimony about distant lights, settled in 1922, in a government circular, about a ridge in North Carolina.

It has a villain, too, of the mild newspaper variety. Col. Wade H. Harris, editor of the Charlotte Daily Observer, wrote to Senator Furnifold Simmons on October 2, 1921 insisting that "a locomotive headlight is easily and unmistakably distinguished as such, not only by the rays it shoots forth, but from its movements," and repeated in his own paper on March 27, 1922 that "a headlight has characteristics that distinctly and unmistakably identify it as such." Simmons and Senator Overman prevailed upon the Survey; Mansfield went; and Mansfield reported that his observations, made in company with persons long familiar with the light, "failed to substantiate these statements." Against Harris he set a woman at Blowing Rock, who told him that on a clear evening you could go out on the hill and see lights popping out all over the valley, "all looking as much alike as so many peas in a pod."

The famous rebuttal — that the lights kept appearing during the great flood of 1916, when no trains ran for weeks — was raised and answered inside that same report. It was a fair hit on the first USGS explanation: D. B. Sterrett, sent in 1913, had "declared that the lights were nothing but locomotive headlights." It was never a hit on Mansfield's, which named four ordinary sources at once. "Automobiles were then in use in the larger towns and on some of the intervening roads, and their headlights were doubtless visible from Loven's over Brown Mountain." That is inference from the roads of the period rather than measurement taken during the flood, and it has stood for a century anyway. The argument is still recited as though unanswered.

Two details finish the case. Southern Railway's engineering department told Mansfield in 1922 that the headlights then in common use on the system were "incandescent nitrogen lamps rated at 250 watts and 32 volts. Fitted with 16-inch silvered-copper parabolic reflectors, these lights yield about 600,000 candlepower." His point was reach, not glamour: the Minots Ledge light was rated at 75,000 candlepower and visible past twenty-five miles, so a headlight could carry the forty-five miles from Blowing Rock to Thermal City. It was brighter than some harbor lights and nothing beside the great coastal beacons — Navesink burned at twenty-five million. Powerful electric locomotive headlights, he wrote, "began about 1909." And the first published account of the Brown Mountain light is a dispatch from Linville Falls to the Charlotte Daily Observer dated September 23, 1913.

The ancient pedigree everyone cites arrived later, and can be dated. Wade Speer's 2017 study finds the first published claim that the lights figure in any Native American tradition in an Asheville Citizen article of 1938, asserted as fact with no source; experts on Native traditions call it a myth invented by white writers. The first ghost story to incorporate the lights appears in 1936. A Revolutionary War soldier arrives in 1982. A Civil War ghost arrives in 2012, on the internet. The often-repeated 1771 de Brahm sighting is false; he never reached the area. Appalachian State ran low-light cameras over the ridge for 6,300 hours through 2014 and recorded nothing unexplainable. Joe Nickell's verdict is the correct one, and it generalizes: "There is no single explanation because there is no single phenomenon."

What the instruments found

• Brown Mountain, North Carolina. Mansfield's plane table and telescopic alidade, 1922; Appalachian State's low-light cameras, 6,300 hours through 2014. Twenty-three lights plotted: 47 percent automobile headlights, 33 percent locomotive, 10 percent stationary, 10 percent brush fires. Nothing unexplainable on camera. DEBUNKED — within Mansfield's own scope.
• Marfa, Texas. The University of Texas at Dallas Society of Physics Students, four nights, May 2004; Stephan, Ghimire, Stapleton and Bunnell in American Journal of Physics 77(8), 2009, twenty nights in May 2008. Every light observed over four nights southwest of the view park attributable to headlights on U.S. 67; spectroscopy that pins a light's distance to 6 percent using atmospheric oxygen absorption. DEBUNKED, with a conceded residue.
• Paulding, Michigan. Jeremy Bos and the Michigan Tech student chapter of SPIE, October 2010; Seal, Lombardi-Rice and Rice, fieldwork October 1990, published in Wisconsin Academy Review 40(2), 1994. Headlights on US-45, about 4.5 miles off, under a stable inversion. DEBUNKED, three times.
• Min Min, Queensland. J. D. Pettigrew, Clinical and Experimental Optometry 86(2), March 2003. A Fata Morgana — refracted light from sources beyond the horizon. DEBUNKED
• Hessdalen, Norway. Project Hessdalen since 1983; the Blue Box station running continuously since 1998; the EMBLA campaigns; a VLF survey in the Journal of Applied Geophysics, 2024. Unexplained. Several serious mechanisms on the table, none confirmed — and no status label that fits, which is itself the finding.
• Ball lightning. Cen, Yuan and Xue, two slitless spectrographs, Tibetan Plateau, July 2012; Physical Review Letters 112(3), 17 January 2014. Real. The spectrum shows silicon, calcium and iron — vaporized soil. DOCUMENTED

The Marfa demonstration is the one to keep. The Dallas physics students parked a car on U.S. 67, flashed the headlights, and produced a Marfa light at the viewing park; a car overtaking that vehicle appeared as one Marfa light passing another. Light frequency tracked traffic volume. Motion ran in a straight line matching the highway. The Texas State team that followed logged nothing over twenty nights but streetlamps, headlamps and small fires — and then conceded that genuinely anomalous Marfa lights are rare and unstudied, which is the honest shape of a debunking. Meanwhile the founding legend has no paper at all: the cowhand Robert Reed Ellison and his 1883 Apache campfire have no contemporary documentation, the Handbook of Texas entry's earliest cited source is a 1976 folklore collection, and the first published account of the lights is Paul Moran's Coronet article of July 1957.

Paulding is the demonstration with a punchline. Michigan Tech students looked through a telescope at the ghostly brakeman's lantern, saw cars, noticed an Adopt-a-Highway sign in the same field of view, found it on a map, and drove a team member out to it. Every appearance of the light matched a passing car. When he switched on his hazards, the phantom blinked yellow. The multicolored displays are police flashers; the rising and falling is a hill; the drifting is the absence of any frame of reference. The first recorded sighting dates from 1966, shortly after that stretch of highway was rerouted. And it had been solved twice already — by a paranormal investigation team in the 1980s, and by a Wisconsin scientific team in 1990. Nobody noticed either time. In 2017 the Science Channel asked Jeremy Bos to re-create his investigation for a mystery series. Ben Radford has the sociology right:

"it's more fun to imagine the distant glimmer is a ghostly railroad brakeman's phantom lantern than the headlights of a 2005 Honda Civic."

Jack Pettigrew explained the Min Min light by driving ten kilometers away on a cool evening after a hot day and dropping his car into a watercourse below the line of sight. Six independent observers watched his headlights float above the horizon. A temperature inversion turns the air into an optical fiber, carrying light from sources far beyond the horizon and lifting them into the sky. The explanation even accounts for a documented shift in the reports: older Aboriginal accounts describe stationary lights, consistent with refracted campfires; modern accounts describe moving ones. He published in an optometry journal, because half the answer is a single isolated light source with no reference landmarks — which is to say, half the answer is in the eye.

Two that stayed open

Hessdalen is the real thing, and it should be said plainly. A standing instrumentation campaign in a Norwegian valley since 1983; a 1984 field report logging 53 anomalous observations across cameras, radar and magnetometers; the Blue Box, a sensor-equipped shipping container, recording continuously since 1998; publications in Acta Astronautica and, in 2024, the Journal of Applied Geophysics. Nobody has explained it. Serious mechanisms are on the table: radon-driven dusty plasma with Coulomb crystals, and a valley-scale geological battery with the river Hesja as electrolyte, which Jader Monari's team tested by drawing enough current between rocks on opposite banks to light a bulb. Nothing is confirmed. The mundane candidates — landing lights on the Oslo–Trondheim route, misidentified stars and aircraft, ground-level lights visible in the station's own photographs — have not been systematically excluded, an objection raised by Matteo Leone from inside the Italian committee. And the project's own sensor list runs: night-color cameras, a passive-radar SDR, an infrasound sensor, and a random event generator. One of those is not an atmospheric instrument. Its 2026 conference program runs from sensor data through to "first-hand testimonies of contactees." That single line is the honest portrait of the field — and the founding posture is still the right one. Erling Strand, 1984: whatever it is, can be measured.

Ball lightning graduated. On a July night in 2012 on the Tibetan Plateau, Chinese researchers studying ordinary cloud-to-ground lightning caught one by chance at 900 meters on two slitless spectrographs: 1.64 seconds of video from formation to decay, plus 0.78 seconds at 3,000 frames per second. About five meters across, traveling fifteen meters at 8.6 meters per second, its brightness oscillating at 100 Hz — probably driven by a nearby 50 Hz transmission line. The spectrum showed neutral silicon, calcium, iron, nitrogen and oxygen. Vaporized ground, matching a silicon-nanoparticle hypothesis published in Nature in 2000. A thousand years of anecdote — beginning with the entry in Gervase of Canterbury's chronicle for June 1195, which modern readers identify as ball lightning only in retrospect — ended in a measurement. The answer was dirt. And it still rests on a single event, which is worth saying too.

The optics of the orb

The photographic orb is not a mystery at all; it is three ordinary facts stacked. A dust mote, pollen grain, insect or raindrop sits inches from the lens, far inside the near limit of depth of field, so it renders as a defocused disc. The inverse-square law makes it vastly brighter than the subject twenty feet away. And in a compact camera the flash sits close to the lens axis, so the particle's illuminated face is precisely the face the lens sees. Fujifilm published this as an ordinary photography problem, because that is what it is.

The decisive test takes one move: put the flash on an arm, off the lens axis, the way underwater photographers have done for decades. The orbs disappear. The reason televised ghost hunting is an orb factory is that infrared night-vision rigs mount superbright IR LEDs directly beside the lens — the worst possible geometry, adopted for reasons that have nothing to do with spirits.

There is a book-length case for the other reading, Heinemann and Ledwith's The Orb Project of 2007, and it deserves naming rather than mocking. But the durable finding here is lexical, not photographic. Annette Hill, writing in 2010, put it exactly: with digital photography, "the ghost light is re-imagined as an orb."

The vocabulary migration

Watch the word travel. Ignis fatuus sounds ancient and isn't: the Latin tag is early-modern, not classical, and is not attested in antiquity. It was applied to marsh flames like the ones Major Blesson chased across a Prussian bog in 1832 and found fled from the air he moved. Screening his breath, he lit a piece of paper at one. The chemistry stayed stubborn for 190 years, until Zare's group demonstrated "microlightning" discharges between rising methane microbubbles in PNAS in September 2025, supplying the ignition mechanism the marsh-gas hypothesis had always lacked. Nobody sees will-o'-the-wisps anymore. The ghost didn't die; the fens were drained.

Then the lantern became a saucer, the saucer became a drone, and the drone became an orb — and the orb walked out of the ghost-hunting forums and into the Pentagon. In May 2023 Sean Kirkpatrick told NASA's UAP panel that the military sees "metallic orbs" worldwide and that spheres make up nearly half of AARO's caseload. NASA's own report of September 2023 captions an MQ-9 clip as "an apparent silver, orb-like object." Thirteen years after a media scholar wrote the sentence, the vocabulary of ghost-hunting message boards was the vocabulary of a federal briefing.

New Jersey then ran the entire cycle at speed. First confirmed sighting, November 13, 2024. By mid-December the story had grown to coordinated car-sized craft over installations, reservoirs and nuclear plants, a blocked medevac, a followed Coast Guard vessel, sprayed gray mist. ALLEGED On December 16, four federal agencies — the Department of Defense, DHS, the FAA and the FBI — reported that more than five thousand sightings had produced nothing anomalous. The Coast Guard swarm was commercial jets making S-turn approaches into JFK. The gray mist was a Beechcraft Baron trailing condensation in turbulence. The viral video posted by the former governor of Maryland was the constellation Orion, per the director of the University of Maryland Observatory, who had to say so out loud. Over one weekend the NYPD took about 120 drone calls, more than in the entire preceding month, many of them aircraft, meteor showers, or Venus — and hobbyists launched drones to look for the drones, generating more reports. On January 28, 2025 the White House concluded they were mainly FAA-authorized research drones and hobbyists, against a national baseline of nearly 2.8 million drones flying in 2024. Sighting, legend, instrument, deflation: seventy-six days, start to finish.

The counterweight

None of this licenses contempt, and the file contains the proof. Reports of enormous flashes above thunderstorms go back to at least 1886, and pilots kept filing them. C. T. R. Wilson predicted upper-atmosphere electrical breakdown on theoretical grounds in 1925 and may have seen one himself in 1956. Atmospheric science had no place for them and treated a century of reports as misperception — until July 6, 1989, when University of Minnesota researchers testing a low-light video camera caught one by accident. Sprites are now imaged from the ground, from aircraft, and from the International Space Station, filmed at 100,000 frames per second, and collected by a NASA citizen-science program called Spritacular. Their root cause, beyond an association with positive cloud-to-ground lightning, is still unknown. NASA's 2023 UAP panel names them explicitly as the model, and states the discipline required:

"Science is a process that reveals reality rather than sculpts it — no matter how unsatisfying or confusing that reality might be."

The hurried debunker and the true believer make the same mistake in opposite directions. Both decide before measuring. Mansfield measured and got headlights. Pettigrew measured and got a mirage. Cen and his colleagues measured and got soil. The Norwegians have measured for more than forty years and got an honest we don't know, which is a different and better kind of not-knowing than a season finale.

The lights are real. The lights have always been real. Nobody has ever seriously doubted that a great many sober people have stood in the dark and seen something bright and strange over a ridge, a river, a highway, a swamp. That was never the question. The question is what the experience is — and the answer, over and over, in federal circulars and optometry journals and Physical Review Letters, is that the world is stranger and more ordinary than the legend, at the same time, in the same place.