The Marfa Lights: Headlights, Mirages and a Riddle the Desert Won't Quite Surrender
One evening in 1883, a sixteen-year-old cowhand named Robert Reed Ellison was driving cattle through Paisano Pass, in the high, dry country of far west Texas, when he noticed a flicker of light out in the dusk. He took it for the campfires of Apaches and said so. Riders went out to look. They found no ashes, no embers, no ring of stones, no trace of a camp of any kind. The next night the light came back, and the night after that. By 1885 other settlers in the basin, among them the Humphreys family, were reporting much the same thing. Ellison's sighting is the oldest attached to the phenomenon that would later be called the Marfa Lights, and it carries one important qualification: he never wrote it down. His account reached paper only decades later, carried there by family recollection.
For the seventy years after Ellison, the record is entirely oral: ranchers, soldiers, and stories told at the family table. During the Second World War, pilots training at the nearby Marfa Army Airfield are said to have searched for the lights from the air and come back with nothing. The first known account of the lights to appear in print arrived only in July 1957, in the magazine Coronet. In 1976 the folklorist Elton Miles gathered nineteenth-century recollections and a rancher's photograph into his book Tales of the Big Bend. Both were memory reaching backward rather than contemporary record. No dated nineteenth-century document describing the lights is known to survive.
What witnesses describe is consistent in outline and maddeningly variable in detail. On dark, moonless nights, glowing spheres appear over Mitchell Flat, the wide basin east of town, roughly between a telephone company tower and the distant profile of Chinati Peak. They hover, split apart, merge again and dart. Sometimes they are white, sometimes tinged with orange or red. Sometimes they hang almost still; sometimes they race sideways and wink out. Reports across the decades have ranged from a single steady glow low on the horizon to a cluster of moving orbs, and the witnesses include ranch hands, tourists, sheriffs and physicists alike. The phenomenon does not do the same thing twice.
The sightings became routine enough to be built for. The state of Texas put up an official viewing platform on U.S. Highway 90, about nine miles east of Marfa, complete with parking and restrooms, and every Labor Day weekend the town holds a festival in honor of the lights.
Science has looked at the basin more than once. In May 2004 a group from the Society of Physics Students at the University of Texas at Dallas spent four nights at the site with traffic counters, video cameras, binoculars and a chase car working the roads below. Their finding was blunt. Every light they observed over those nights lined up with vehicles on U.S. Highway 67, which crosses Mitchell Flat on its way south to Presidio. When a car from their own team flashed its headlights out on the highway, the watchers back at the platform saw a Marfa Light appear at exactly that spot. The frequency of the lights rose and fell with the volume of the traffic.
The mechanism the students identified was atmospheric refraction. The desert floor around Marfa, which sits at nearly 4,700 feet, cools sharply after scorching days, and the temperature can swing forty or fifty degrees between afternoon and night. That stacks layers of air of sharply different temperature and density close to the ground, the same physics that produces the shimmering Fata Morgana mirage at sea, and those layers can bend, stretch and lift the light of distant headlights into wobbling, floating globes that appear to hang in mid-air miles from any road.
The most sustained attempt to account for whatever the traffic does not cover belongs to James Bunnell, a retired aerospace engineer who grew up in the Big Bend country and returned to it to study the problem. Beginning around 2000 and continuing for close to a decade, he ran automated monitoring stations around the basin, eventually three of them carrying nine cameras between them, recording night after night, unattended, so that no human expectation could shape what the lenses caught. The overwhelming majority of what he recorded was headlights. But across those years his network logged roughly forty events he was unwilling to file as traffic: lights that moved against the wind, reversed direction abruptly, flared to extraordinary brightness and vanished, or appeared in the frame when his instruments recorded no vehicle anywhere on the flat. In his 2009 book Hunting Marfa Lights he put the unexplained fraction at about three percent.
The striking displays are rare. Sorting through years of automated footage, Bunnell estimated that lights he could not attribute to traffic appeared on average only a handful of times across roughly five nights in a year. This is not a spectacle available on demand. It is something that must be waited for, across a basin many miles wide, on nights that cannot be predicted in advance.
Sharper instruments followed. Karl Stephan, an engineering professor at Texas State University, led a campaign spread over some twenty nights using a Schmidt-Cassegrain telescope coupled to a spectrometer. The method exploited a property of the atmosphere: oxygen in the air absorbs light at a known band of wavelengths, and the depth of that absorption in a distant light's spectrum indicates how much air the light has travelled through, and therefore how far away its source is. Every light his team caught that was bright enough to yield a usable spectrum proved to be a vehicle headlamp, a distant streetlight, or a small brush fire.
Neither Bunnell nor Stephan closed the file. In 2011 the two men jointly published an analysis of an exceptionally bright object recorded on the night of 3 June 2005 by two of Bunnell's stations at the same instant. Because two separated cameras had captured it together, its position could be triangulated. When they worked out where the object actually was, it did not sit on any road, and it did not resolve into traffic. That single event, and the small archive of similar recordings behind it, is where the case stands today.
Conclusions and Open Questions
The strongest explanation on the table is refraction of headlights, proposed in its modern form by the University of Texas at Dallas physics students in 2004 and reinforced by Karl Stephan's spectroscopic work. It accounts for almost everything visitors photograph from the platform, and it has the rare virtue of a successful prediction: the students made a light appear on demand by driving a car onto Highway 67. Its weakness is the word almost. Refraction of headlights requires headlights, and the legend appears to reach back to a decade when there were none in west Texas.
That leaves an awkward chronological fork, and both branches cost something. If Robert Ellison really saw a light at Paisano Pass in 1883, the tidy modern explanation is incomplete, because there was nothing on Mitchell Flat for the atmosphere to bend. If his account was invented or embroidered in the retelling, which is possible given that it reached print only after cars had crossed the basin for decades, then the phenomenon loses the historical pedigree that makes it more than a roadside curiosity. No dated document decides the question either way.
For the residue his cameras caught, Bunnell has raised the possibility of some form of electrical discharge, perhaps a piezoelectric effect produced by pressure on the igneous rock beneath the basin, a cousin of the disputed earthquake lights that some geologists argue accompany seismic strain. The idea names a specific local cause: Mitchell Flat lies in the shadow of the Chinati Mountains, the deeply eroded remains of an ancient volcanic field, and the rock below is igneous. Its weakness is that no such mechanism has been demonstrated to produce free-floating lights at Mitchell Flat or anywhere else, and Stephan, who has considered it closely, has declined to endorse it.
A duller possibility is rarely given its due. Some argue that the leftover lights are also mirages, refractions of exactly the same kind as the common ones, but of sources nobody has bothered to catalogue: a ranch truck on a back track, a campfire, a locomotive headlamp on a distant line, a flare beyond the visible horizon. The attraction of this theory is that it requires no new physics at all. Its weakness is that it is unfalsifiable as stated, since an uncatalogued source is by definition one nobody has found. Nothing documented confirms it, and nothing yet rules it out. At the far edge of the discussion sit claims of exotic or extraterrestrial origin, which one theory holds explain the erratic motion; they remain assertions rather than findings, since no instrument at Marfa has ever recorded anything requiring them.
What remains unexplained is small and specific: roughly forty logged events, about three percent of Bunnell's record, including the object of 3 June 2005 that two cameras placed away from any road. The open questions follow from that. Is Marfa a nineteenth-century phenomenon or a twentieth-century one wearing an inherited costume? Why should a natural mechanism appear only a few unpredictable nights a year? And can a phenomenon that rare, in a basin that wide, ever be cornered by instruments at all? Ninety-seven percent of the file is solved. Closing the last three percent has so far defeated everyone who has tried.