On the dark horizon near Marfa, Texas, lights have been described as changing colour, dividing and vanishing. The case contains several different kinds of evidence: remembered nineteenth-century sightings, a newspaper report from 1940, controlled tests of distant headlights, and later recordings made during thunderstorms. Those records deserve separate examination. An explanation demonstrated for a particular light does not identify an earlier witness’s light; an unidentified recording does not establish a new physical phenomenon.
The landscape and the remembered beginnings
Julia Cauble Smith’s Handbook of Texas entry places the tradition between Marfa and Paisano Pass, facing the Chinati Mountains in northeastern Presidio County. It recounts Robert Reed Ellison’s supposed 1883 sighting and the Humphreys family’s 1885 experience. These are historical recollections in a later reference work, rather than contemporary documents reproduced there. The entry also records unsuccessful searches for a source and discusses a mirage-like atmospheric explanation. Its account establishes the local tradition; it does not independently date each remembered experience.
A newspaper report already in print in 1940
The front page of The Sanderson Times, 30 August 1940, carries Bill Goldwire’s article about a light seen by motorists near Marfa. Goldwire described an intermittently visible light in distant hills and a succession of white, red, blue, yellow and green appearances. He reported that Border Patrol officers attributed it to vehicles crossing a pass, but argued against that interpretation from the apparent remoteness, persistence and colour changes.
Goldwire also attributed to Dr Monroe Slack a recollection of seeing the light forty years earlier. That is evidence that the recollection was being reported in 1940, rather than a dated record from around 1900. A proposed phosphorus-vapour source was speculation, without chemical sampling. Stories of missing searchers and a wandering old man accompanied the discussion; the article encouraged readers to use their imagination. This juxtaposition of suggested physical causes and storytelling is part of the contemporary record. The issue proves that press coverage existed before 1957, without proving that this was the first published account or identifying what Goldwire saw.
The UT Dallas headlight experiment
The Society of Physics Students progress report, dated 10 December 2005, names Alexander Stolyarov, Jeff Klenzing, Patrick Roddy and faculty adviser R. A. Heelis. Its introduction places the expedition on 10–14 May 2004; its conclusion instead refers to observations on 11 and 13 May 2005. That internal year discrepancy should remain visible when citing it.
The group used video, a traffic counter on US 67 and a chase vehicle communicating by radio. Speeded-up recordings revealed regular paths and recurring appearance points. Traffic diminished through the night. More directly, observers recognised the chase car’s flashing headlights in the same horizon area as the supposed mystery lights. Another light crossed its position while the driver reported a passing car. The report attributed all reliably observed lights in its experiment to vehicles. The chase-car test matters because it connected a known source with its distant appearance. Traffic correlation alone would have been weaker evidence. Its conclusion applies to the observations investigated, rather than every story attached to Marfa.
Twenty nights of spectroscopy in 2008
Karl D. Stephan, Sagar Ghimire, William A. Stapleton and James Bunnell reported a different investigation in the American Journal of Physics in 2009. They worked for twenty nights in May and June 2008, combining a telescope, video and a CCD spectrometer. No observed object met their criteria for a genuine Marfa light. Identified sources included vehicle lights and fires.
Their measurements addressed the difficulty of recognising distant sources. Atmospheric disturbance could merge headlights into an unstable patch; unusually clear conditions resolved individual lamps much farther away. Oxygen absorption in the spectrum helped estimate distances, checked against road positions. Fire spectra supplied another means of identification. The paper distinguished those results from Bunnell’s longer camera-monitoring programme, whose selected unusual incidents were relatively infrequent. This was a bounded campaign with useful positive identifications, not proof that every reported occurrence had been observed. Conversely, the authors’ category of “genuine” lights was an investigative selection criterion, not an established class of unknown objects. Their unsuccessful search for such an example cannot establish its physical nature.
An earlier thunderstorm recording, published later
A 2011 paper by Stephan, Bunnell, John Klier and Laurence Komala-Noor examined recordings from 2–3 June 2005, during thunderstorms. One especially bright source persisted for more than three hours. Photometry and triangulation suggested a distant source with visible output around ten kilowatts at its peak. That estimate depended on assumptions about spectrum and radiation; the location had substantial uncertainty. Exact monitoring-site coordinates were withheld at landowners’ request.
A power line crossed the possible location area. A utility manager considered an electrical arc possible; the authors argued that duration, gradual brightening and apparent movement made that explanation less satisfactory. They considered combustion and several more speculative mechanisms, without establishing one. The paper explicitly left uncertain whether this event shared a mechanism with other Marfa reports. It should therefore be distinguished from the later twenty-night campaign, and from a confirmed identification of ball lightning. The estimated power is a model-dependent result, not a direct measurement of an unknown object’s total energy. Disagreement over an electrical explanation remains part of the interpretation.
What the investigations leave open
The research programme had already been discussed in Stephan’s 2006 American Physical Society meeting abstract. It considered headlights, mirages, natural gas and more exotic proposals, including arguments against antimatter and evaporating black holes. This is a conference abstract, not a full experimental report. It helps document the questions investigators were asking, but cannot substitute for the later methods and data.
Together, the records support a careful distinction between observation and identification. The controlled vehicle experiment supplies a demonstrated source for some lights. The spectroscopy supplies additional source discrimination. The thunderstorm recordings contain measurements whose interpretation remained disputed. Historical recollections supply another category again. Combining all four into a single count of unexplained objects would erase differences in what was actually recorded. Equally, assigning every recollection to the tested cars would go beyond the evidence. A stronger future report would connect a timed observation to simultaneous independent measurements and a recoverable source location, while retaining the original recordings and uncertainty estimates.
The public viewing place
Visit Marfa identifies the official viewing area nine miles east of town on US 90 towards Alpine. Its account promotes the lights as a visitor attraction and describes contrasting interpretations. That is useful information about the public setting, rather than an independent experimental verdict. The map pin here identifies the viewing platform, using the photograph’s documented camera location. It does not identify a light’s origin, a private ranch or a research station. The red lamps visible in the accompanying photograph belong to the visitor area.
Sources and related cases
The 1940 newspaper page and the three-page UT Dallas report were read in full for the relevant account; the report’s conflicting years are retained above. The 2009 paper was consulted through its author-uploaded full text, and the 2011 study through its accepted-manuscript copy. These are distinct studies of different observations. For another historical investigation of distant lights, see Brown Mountain and Mansfield’s 1922 fieldwork. For claims involving vehicles in Texas, see the Levelland reports of 1957; they are a separate case, with different witnesses and records.