The Levelland UFO, 1957
Stalled Engines on a Rainy Texas Night, and What the Air Force Actually Concluded
On the night of 2 to 3 November 1957, drivers around Levelland, Texas reported engines dying and lights going out as a glowing object passed. The Air Force sent an investigator, interviewed six people, checked the weather, the oil fields and the air traffic, and could not explain it. The squadron's report is in the Pentagon's eighth UFO file release. We read all 21 pages.
| Dates | Night of 2 to 3 November 1957 (five sightings), plus one on 4 November |
| Place | Roads around Levelland, Hockley County, Texas |
| What was reported | Engines and lights failing as a glowing oval or 'torpedo' passed; streaks of red light |
| Weather | Complete overcast, 400 ft ceiling, visibility 3 miles, drizzle all night (Air Force check) |
| Investigators | 1006th Air Intelligence Service Squadron, Ent AFB, Colorado; interviews 4 to 6 November |
| Air Force verdict | 'Forwarded as unresolved' (report AISS-UFOB-386-57, 8 November 1957) |
| Pentagon file | DOW-UAP-D191, PURSUE Release 8, 20 November 2026 |
The Levelland UFO is one of the best-known car-stalling cases of the 1950s. Over a few hours on a wet Saturday night in November 1957, people driving the roads around Levelland in west Texas reported to the police and the sheriff that their engines had died and their lights had gone out as something glowing passed them. It came a few hours after the Soviet Union launched Sputnik 2, and the story went national.
The Pentagon's eighth UFO file release includes the Air Force's own investigation, file DOW-UAP-D191. The government's summary doesn't say it's Levelland, but it is: 21 pages from the 1006th Air Intelligence Service Squadron, with six interviews, three memory sketches, a hand-drawn map and a verdict that surprises people who know only the official story. The squadron didn't call it lightning. It wrote that it couldn't resolve it. We read every page, then checked the sky for that night with Skyfield and NASA's JPL ephemeris, the tools we use to verify every sky claim on the site, including our UK night sky this week guide.
What Happened at Levelland, Texas in 1957?
At Levelland in November 1957, two drivers told Air Force investigators that their vehicles died as a glowing object passed, and four other people described lights in the sky over the next two nights. The first sighting in the file came at 11pm Central time on 2 November, about four miles from Levelland on State Highway 116. A truck driver saw what he first took for lightning, "a big flame" ahead and to his right. As it reached him his engine and lights went out. He got out, and "it was so rapid and quite some heat, that I had to hit the ground". He described a blue torpedo about 200 ft long with yellow flame and white smoke at the back, and said it "disappeared by burning out". Once it had gone, his truck started again.
The account the Air Force rated most highly came an hour later. At about five past midnight on 3 November, a 19-year-old Texas Tech student was driving home from Lubbock on the same highway, about four miles west of Smyer. His ammeter jumped to discharge, the engine began "cutting out as if it were not getting any gas", and then the lights and radio failed. He got out, checked the battery, turned round and saw the object.
"SOURCE described the object as oval shaped and he thought that the size of the object was that of a baseball at arm's length. He estimated that the object was seventy-five to one hundred feet at its longest dimension."· Investigator's summary of the student's account (DOW-UAP-D191, p5)
It was white with a greenish tint, which he thought might be his tinted windscreen. He watched it from inside the car for four or five minutes "when it rose virtually straight up and out of his line of vision". His lights and radio came back and the engine started normally. He told nobody until his parents came home from a weekend away, "for fear of public redicule" (his spelling), and reported it to the sheriff that afternoon.
Levelland UFO Witnesses: Who Did the Air Force Interview?
The Air Force interviewed six Levelland UFO witnesses between 4 and 6 November 1957: a student, a truck driver, the county sheriff, a highway patrolman, an Air Force technician and a drilling superintendent. A staff sergeant from the 1006th AISS at Ent Air Force Base, Colorado, did the interviews. The file names all six and gives their home addresses; we describe them by role.
The student
Engine, lights and radio failed; watched an oval object on or over the road for four or five minutes until it rose straight up. Rated 'usually reliable'. The preparing officer called his the most important sighting.
The truck driver
Engine and lights died as a flaming 'torpedo' passed over; he lay on the ground because of the heat and speed. Rated 'not usually reliable', and the file has him heading west in one place and 'north and west' in his own statement.
The sheriff
Out searching for the reported object, he saw 'just a streak of light one time', reddish, crossing from south to west in about two seconds.
The highway patrolman
A red to orange-red flash, east to west, a mile or so ahead and close to the ground, lasting 'only a fraction of a second'.
The Air Force technician
Saw lightning and heard radio static, then a flash in the south-west; his radio and car lights went out for about one to three seconds. He came forward to help explain the others.
The drilling superintendent
Two nights later: a red light swinging north to south under a 'balloon' shape, throwing sparks 'similar to a cutting torch', then shooting up into cloud. The officer judged it unconnected.
The file is frank about how the Air Force graded people. Two witnesses are written off as of "below average intelligence", which says more about 1957 than about the sighting. It's also frank about the gaps. The newspapers reported more witnesses than these six, and the investigator went looking for them. His note: they "had either disappeared or returned to their homes, leaving no forwarding addresses."
Did the Levelland UFO Stop Car Engines? The Lightning Witness
The Levelland UFO file has two drivers whose engines died, and one more whose lights and radio blinked out during lightning, which is the Air Force's best clue and is rarely quoted. That third man was an Air Force aero-medical technician, driving home near Shallowater at about 11:35pm (the file's time) on 2 November. He'd seen occasional lightning and heard static on his radio. Then came a flash to the south-west, and his radio and headlights went out for one to three seconds before coming back.
"My wife and I remarked that was certainly a strong bolt of lightning to put out our lights and radio."
· The Air Force technician's statement (DOW-UAP-D191, p14)He hadn't thought much of it until he read about the other reports, and he came forward to help. The preparing officer used his account as a control: "The fact that the radio and lights were cut out for a few seconds by the lightning in this sighting indicates that lightning could have entered into the other sightings."
That is as far as the file takes it. The student's car was dead for several minutes, not seconds, and it came back to life when the object left. The squadron's own astronomy expert, a professor of mathematics and astronomy at Texas Technological College, agreed lightning could stall a car, "however, the possibility decreases as the number of such incidents increase." The squadron also checked for fallen power lines and found none.
Was the Levelland UFO a Star, Planet or Satellite? Our Sky Check
The Levelland UFO could not have been a star, a planet or a satellite, because the Air Force's own weather check found a complete overcast with a 400 ft ceiling and drizzle all night. That is the most useful single fact in the file. The squadron got the conditions from Reese Air Force Base and the US Weather Bureau at Lubbock: a shallow inversion up to about 6,000 ft, "a complete overcast" with "a 400 foot observed ceiling", visibility three miles, and "a drizzle or light rain throughout the period". The witnesses say the same: heavy cloud, light rain, no stars.
We still ran the numbers, because that's what we do with every case on this site. Using Skyfield and the JPL DE421 ephemeris for Levelland (33.6°N, 102.4°W) on Central Standard Time, our calculation gives:
- 11pm on 2 November: the Moon 49° up in the south-west, 83% lit. Venus had set hours earlier.
- 12:05am on 3 November: the Moon 39° up, a little further west.
- 1:30am: the Moon 24° up in the west-south-west. No bright planet was above the horizon during any of the sightings.
- 8:45pm on 4 November: the Moon 52° up in the south-east, 94% lit, again behind "thick, heavy clouds".
It was a waxing gibbous Moon, a few days short of full (you can check tonight's phase on our is it a full moon tonight page). A Moon that bright, high up and behind thin cloud, can light a patch of sky. Behind a 400 ft overcast with drizzle it can't produce a red streak, a flaming torpedo or an oval on the road. Nothing astronomical explains Levelland, and the file never suggests anything that is. If you ever want to know what's actually up on a given night, our what you can see in the sky tonight page does the same calculation for the UK.
What about Sputnik? Sputnik 2, carrying the dog Laika, was launched at about 02:30 UTC on 3 November 1957 (outside context, not in the file). That's about 8:30pm on 2 November in Texas, roughly two and a half hours before the first sighting here, which is why newspapers linked the two. Without historic orbital elements we can't say where Sputnik 2 or its rocket were that night, but it doesn't matter: nothing in orbit can be seen through a 400 ft overcast. For observing purposes the timing is a coincidence.
What Did the Air Force Conclude About Levelland? "Forwarded as Unresolved"
The Air Force squadron that investigated Levelland concluded that it could not resolve the sightings, and sent the case up the chain marked "unresolved". The last page of the report, the preparing officer's comments, goes through the witnesses one by one. He set aside the truck driver ("cannot be relied upon": 800mph at 300 ft away doesn't fit two to three minutes in sight), linked nothing to the sheriff's and patrolman's brief streaks, and judged the drilling superintendent's swinging light two nights later unconnected. He used the technician's lightning as a possible explanation for the rest. And he gave the student the most weight: "The reliability, sincerity, and intelligence of the SOURCE lend credence to this account of his sighting."
He then listed four possibilities: weather phenomena such as St Elmo's fire; ball lightning, noting that "there were Nimbus clouds in the area, which, according to information at this organization must be present for ball lightning to occur"; "reflection of excess burning gas from a very low cloud cover"; or some combination of those. The squadron had checked the gas: local oil companies were burning "a limited amount of excess gas", and most went back into the ground. It had also ruled out aircraft with four flight services and control points, and weather balloons, which Amarillo and Midland released at fixed times and were "not likely to be in the area".
"It is felt that insufficient knowledge of the above phenomena exists at this organization to completely resolve this sighting. Therefore, this sighting is forwarded as unresolved."· Comments of the Preparing Officer, 1006th AISS (DOW-UAP-D191, p21)
The answer most people know came later and from higher up. Project Blue Book's public explanation for Levelland was a severe electrical storm, most probably ball lightning or St Elmo's fire (outside context; that conclusion is not in this file). The squadron that actually went to Texas named ball lightning as one candidate among several and said plainly it couldn't decide. The file also has its own inconsistencies: the summary page says "two witnesses" saw the object on or near the ground, while the officer's comments say only one did, and the summary's "two hundred feet long" is the truck driver's figure, not the student's.
CONFIRMED
Not astronomicalA complete overcast at 400 ft with drizzle all night. The Moon was up but hidden; no bright planet was above the horizon between 11pm and 1:30am.
CONFIRMED
No aircraft or balloonsFour flight services and control points reported no traffic; the nearest weather balloons were judged unlikely to be in the area.
PLAUSIBLE
Lightning affecting carsOne witness's lights and radio blinked out for seconds in lightning. It doesn't account for minutes-long failures that ended when the object left.
POSSIBLE
Ball lightning, St Elmo's fire, gas flaresAll listed by the squadron; none chosen. Nimbus cloud was present and some gas was being burned locally.
UNRESOLVED
The student's ovalThe squadron's most credible witness, a dead car for four or five minutes, and an object that rose straight up. The file has no answer.
Scan the skies yourself
Kit we've reviewed in full
Nobody at Levelland could have used binoculars on a 400 ft overcast. On a clear night, though, a good pair settles most strange lights in seconds: aircraft show navigation lights, planets hold steady and satellites drift.
Bright, easy to hold steady and quick to get on target, which is what you need when a light is moving.
Pulls in faint, distant lights and gathers plenty of light after dark. Use a tripod at 15×.
Crisp, contrasty images that help you tell a point of light from something with shape.
Affiliate links: you pay the same price, we earn a small commission that helps keep WatchTheStars free.
Orogrande, New Mexico: The Other Car-Stalling UFO That Week
The Orogrande UFO of 4 November 1957 is the other car-stalling report in the same release, and the same Air Force squadron investigated it (file DOW-UAP-D194, report AISS-UFOB-438-57). This one was in daylight. At about 1:10pm Mountain time, a civilian electronics technician from the Air Force Missile Development Center at Holloman was driving south on US 54, eight to ten miles below Orogrande. His radio faded and his car slowed. He joined about six other motorists who had pulled over and watched a pearly, egg-shaped object make two passes across the valley from the Sacramento Mountains towards the San Andres.
He said that as it passed he felt "a distinct rise in his body temperature, as though he had been exposed to a 'giant sun lamp'", and he said he had found a severe sunburn when he got home. The file adds: "(This statement was later retracted)". He also said his car "started again quite easily, although the battery had been steaming." The investigator didn't think much of him as a witness, and the other motorists couldn't be traced. The verdict was a possible optical phenomenon: "Due to the flat white sandy terrain, mirages cannot be considered unusual occurrences."
The Air Force was right that nothing astronomical fits. Our calculation puts the Sun 38° up in the south-south-west and Venus 25° up in the south-south-east, in daylight, with nothing in the north-east where he first saw the object at 35° up. We'd add one thing the file doesn't: mirages hug the horizon, and an object 35° up is a long way from the horizon. Orogrande is the weaker case, but it shows car-stalling reports were coming in from more than one place that week.
How the Air Force Reasoned That Week: White Sands and the Wrong Venus
The White Sands UFO file from the same week (DOW-UAP-D193) shows the Air Force blaming Venus for a sighting made after Venus had set, which is a useful check on how much weight to give its quick explanations. On 3 November 1957, two Army military police patrols at White Sands Proving Ground reported lights near Stallion Site and the Trinity site. The same squadron decided the first patrol, at about 2:30am, had seen moonset, and the second, at 8pm, "merely witnessed the planet Venus". Its own figures put Venus "at 15 degrees elevation, with an azimuth of 225 degrees" at 8pm.
| Air Force, 20:00 MST | Venus 15° up at azimuth 225° |
| Our calculation, 18:15 MST | Venus 14.6° up at azimuth 223°, which matches the Air Force figures |
| Our calculation, Venus sets | 19:49 MST |
| Our calculation, 20:00 MST | Venus 2.4° below the horizon |
| Patrol 1 moonset | Air Force 02:29 MST; ours about 02:50 MST, same direction (west) |
So the Air Force's Venus position is correct for about a quarter past six, not for 8pm, when Venus had already gone down. The moonset explanation for the first patrol is a fair fit, even if its time is about 20 minutes early by our calculation. The wider story of how often the 1950s Air Force reached for the wrong planet is in our post on when the Air Force blamed the wrong planet. The point for Levelland is a simple one. In the same week and the same squadron, a confident astronomical answer could be wrong, but at Levelland the squadron didn't offer one. With a 400 ft cloud base, it couldn't.
Levelland UFO: Our Verdict
The Levelland UFO remains unexplained, and the Air Force's own investigators said so. We're not going to claim a cause either. Here's what the file supports:
- It wasn't a star, planet, satellite or the Moon. The cloud base was 400 ft.
- Lightning was around, and one witness's lights and radio blinked out for seconds in it. That makes some weather explanation for the brief streaks and flashes reasonable.
- The two engine failures are the hard part. The student's account, which the squadron rated highest, describes several minutes without power that ended when the object left, and weather doesn't obviously do that.
- Most of the witnesses the newspapers described were never interviewed, so the file holds six accounts, not the whole night.
That sets Levelland apart. Most of the 1950s Air Force reports we've checked against the sky turn out to be a planet, a bright star or a meteor, and 1957 has its share of odd military reports too, such as the B-52 radar contact on the round-the-world flight. Levelland doesn't reduce to anything in the sky.
It belongs with the other cases in our UAP research hub as a genuine open question, honestly recorded at the time. And it's a good reminder of what a single number can do. "400 foot observed ceiling" rules out more explanations than pages of argument. If you see something odd yourself, note the time, the direction and the weather, check ISS passes tonight, and look at our guide to why a star seems to follow your car or plane. If you're new to the sky, start with stargazing for beginners.
Sputnik 2 launched
About 02:30 UTC on 3 November, which is 8:30pm on 2 November in Texas (outside context).
The truck driver
Engine and lights fail on Highway 116 as a flaming 'torpedo' passes. 23:35: the Air Force technician's lights blink out in lightning near Shallowater.
The student
3 November: car dies west of Smyer; an oval object sits on or over the road for four or five minutes, then rises straight up.
Sheriff and patrolman
Searching the Oklahoma Flat road, they see brief red streaks.
Sundown, and Orogrande
A swinging red light near Sundown at 20:45 CST; earlier that day, the Orogrande car-stalling report in New Mexico.
Report AISS-UFOB-386-57
Interviews done 4 to 6 November. The squadron forwards the case as unresolved; Headquarters USAF logs it on 18 November.
Published in full
20 November: the report appears in PURSUE Release 8 as DOW-UAP-D191.
Primary Source Documents
The Levelland file, 1957
DOW-UAP-D191 · 21 pages · PURSUE Release 8
The 1006th AISS report: six interviews, three sketches, the vicinity map, the weather and traffic checks and the verdict.
Download PDFOrogrande, 4 November 1957
DOW-UAP-D194 · 10 pages · PURSUE Release 8
The daylight car-stalling report on US 54, the retracted sunburn and the mirage verdict.
Download PDFWhite Sands, 3 November 1957
DOW-UAP-D193 · 8 pages · PURSUE Release 8
Two military police patrols, a moonset and a Venus that had already set.
Download PDFNote: Official US Department of War records released through PURSUE (Release 8, 20 November 2026, war.gov/ufo), mirrored on our own servers. Every record from all eight releases is in our Pentagon UFO file browser, and our guide to the PURSUE releases explains what each batch contains. Sky positions are our own calculation (Skyfield, JPL DE421, historical time zones). Witnesses and junior personnel are described by role; their names and addresses in the file are not reproduced. Sputnik 2's launch time and Project Blue Book's later explanation are outside context, not from these files.
Related Cases
More UAP cases
- The RB-47 Encounter: 700 Miles, Three Sensors, No ExplanationMississippi → Texas → Oklahoma → Kansas · 17 July 1957 · ~0200–0340 CST
- Apollo 17's Unknown CompanionTranslunar Coast · 9 December 1972 · Mission Elapsed Time 02:18:47
- RAF Lakenheath: The Night Britain's Air Defences FailedSuffolk, England · 13–14 August 1956 · 2130–0330 hrs
- The 1976 Tehran UFO Incident: F-4 Jets, Weapons Failure and the Declassified DIA Report
- The Senator Russell Sighting4 OCTOBER 1955 · SOVIET TRANSCAUCASUS (PRESENT-DAY AZERBAIJAN) · FULL CIA FILE RELEASED 10 JULY 2026
- The Chupa-Chupa AttacksSeptember–December 1977 · Colares, Vigia & the Pará coast · Status: Documented, Unexplained


