A six-man USAF RB-47H crew reported unusual lights and electronic signals during a flight across the southern United States on 17 July 1957. The account survives in a same-day ground-station message, a wing-intelligence summary and aircraft commander Lewis D. Chase's September questionnaire. Later investigations recovered these records after the Colorado UFO project's unsuccessful archival search. Their significance lies in the claimed relationship between cockpit observations, passive direction-finding signals and ground-radar reports. The surviving documents also contain contradictory statements about radar contact, timing and recordings, which prevent treating the episode as one continuous, independently preserved track.
Chase and McCoid in the cockpit, Hanley navigating, and Provenzano, McClure and Tuchscherer at electronic-monitor stations
The flight in the 1957 records
Chase's questionnaire identifies him as an RB-47H aircraft commander with the 38th Strategic Reconnaissance Squadron at Forbes Air Force Base. McDonald's later account names the remaining officers and stations: James H. McCoid, co-pilot; Thomas H. Hanley, navigator; John J. Provenzano, monitor one; Frank B. McClure, monitor two; and Walter A. Tuchscherer, monitor three. The early form lists the commander and co-pilot as visual observers. The electronic-warfare officers worked in the aft compartment and supplied a different kind of evidence.
The four-page wing-intelligence report prepared by Elwin T. Piwetz describes McClure's monitor intercepting a signal near 3,000 MHz and its bearing moving unusually across the display. It records a frequency of 2,995–3,000 MHz, two-microsecond pulses, 600 pulses per second, a four-revolutions-per-minute sweep and vertical polarisation. These are reported signal characteristics in a narrative summary. Direction finding indicates the apparent arrival bearing of a transmission; it does not independently measure the distance to a solid object.
At 10.10 am UTC, Piwetz's report places an intense white light with a blue tint crossing ahead of the aircraft from about eleven o'clock towards two-thirty. It attributes that observation to Chase and McCoid. A subsequent electronic check is recorded at about 10.30, followed by another visual observation at 10.39 and a report of two red lights and two signal bearings at 10.40. Chase's own form describes intermittent observations over about an hour and a half and repeatedly states that only light, without determined size or dimensions, was visible.
The crew obtained permission to depart from the flight plan and sought assistance from the ground-control station using the call sign Utah, identified in the later reconstruction as the Duncanville station in Texas. Piwetz reports ground confirmation, a loss of contact near 10.50 and reacquisition during the aircraft's turn. At 10.57, however, the report records an electronic bearing while Utah had no scope contact. It later describes the signal behind the northbound aircraft, fading near Oklahoma City at 11.40. The report's 11.02 departure time differs from the 11.20 transcription in McDonald's article.
Chase completed the Airborne Observer's Data Sheet on 10 September 1957. He marks the attempt to detect the phenomenon on airborne radar unsuccessful and separately reports that Utah tracked it. His conclusion that the source was solid relied on that ground report, rather than a resolved visual structure. An August request for the questionnaire and the 17 October forwarding endorsement document the reporting chain; the latter carries the Air Defense Command's then-unexplained assessment.
Later interviews changed some details. McDonald notes that no interviewed crew member recalled the early report's occasional second source, and distinguishes their immediate electronic-recheck recollection from the twenty-minute interval in Piwetz's chronology. Klass subsequently reproduced favourable responses from Chase and McClure to his conventional reconstruction. Those published selections document a qualified later position, rather than an unchanged collective endorsement of every extraordinary interpretation.
Evidence review
The principal onboard electronic evidence belongs to McClure's passive receiving system. McDonald describes an ALA-6 direction finder, APR-9 receiver and ALA-5 pulse analyser; Provenzano later checked a similar bearing on his own system, while Tuchscherer's monitor covered a lower frequency range and did not detect the signal. This is distinct from Hanley's active navigation radar. Chase's unsuccessful-radar answer was recorded within two months of the flight and offers a record closer to the flight than another officer's interphone recollection a decade later. It records the result of an attempt, without supplying a continuous account that could exclude a return during every later phase. A navigation-radar track remains disputed rather than established across the crew.
Cockpit and electronic observations were discussed through the aircraft interphone. Ground positions were transmitted by radio. The reporting chain can preserve real agreement while also linking how the participants interpreted their displays and lights. Establishing that all channels followed one source requires time-aligned bearings and independently preserved records; the inspected documents supply summaries and testimony rather than continuous signal waveforms, ground-radar film or a complete synchronised recording.
Recording claims are inconsistent within the evidence history. Chase circled no object photographs but added that radar-scope pictures had been taken. That note could refer to an electronic-monitor display rather than the active navigation radar and does not identify which instrument was photographed. Piwetz records monitor three recording interphone and command conversations, not a continuous intercepted-signal waveform. The Colorado study later found the pilot's recollection of seized film and recordings contradicted by the co-pilot and monitor-two officer, who said recording materials were not aboard. No such original recording was inspected here. The conflict leaves both the existence and contents of a retained recording uncertain.
57-page mixed compilation, containing both 1957 records and later additions. Selected scans inspected; page 1 record card, pages 3–4 request/endorsement and pages 5–6 assessment notes used. Page 4 visibly 17 October 1957. The complete compilation was not reviewed page by page.
All twelve questionnaire pages inspected as scans, with non-sequential compilation order. PDF41: failed airborne attempt, reported Utah tracking and scope-picture note. PDF48: only light/no dimensions. PDF52: role and completion date. No original scope pictures inspected.
PDF37–40 fully read as scans. Elwin T. Piwetz identified as wing intelligence officer. Precise report-writing day not established. Source's correlation conclusion attributed; 11.02 timestamp retained as printed, with later 11.20 transcription distinguished.
Three split message pages read at PDF54, 13 and 35 (also reproduced partly on 24/46). Contains aircraft-reported contact and ground-contact limitation; handwritten annotations have unidentified authors and dates and are not adopted as original message content.
All five re-typeset PDF pages read and compared with selected magazine scans in the official compilation. Includes AIAA introduction, map and source-owned reconstruction. Primary October forwarding date overrides its November statement. Original interview tapes and Chase approval letter not inspected.
Full short case text read through NCAS transcription. Retrospective interviews and unsuccessful record search, including recording-history contradictions; predates recovery of the 1957 file. No original interview recordings examined.
Relevant Case 5 section read, NCAS lines 426–477. September weather profiles expressly dated differently from actual July event. Thayer's tentative unknown conclusion and proposed explanations retained with that date limitation.
Selected original paper pages read as scans: PDF15,16,20,23,25,29–31; supporting extracted passages also read. Printed16/PDF31 dated 30 December 1971 includes selected Chase/McClure responses. Original correspondence, complete equipment manuals, maintenance diagnosis and airline arrival records not independently inspected.
Entire PDF57 scan read. Distinguishes Amarillo TWA episode from Flight655/966 encounter near El Paso. Does not supply a full RB-47 route identification or actual Dallas arrival log.
Signal bearings, visual observations and the conflicting radar record
The early records do not support a simple claim of either universal radar confirmation or universal ground-radar failure. The Blue Book record card says the object was detected by B-47 radar but not ground radar. Chase's questionnaire says the airborne attempt failed and ground tracking occurred. Piwetz describes reported ground contacts alternating with losses. The Duncanville message says the aircraft reported good contact while ground radar lacked contact at one stage. These are conflicting and phase-dependent statements, not interchangeable summaries of a single technical result.
Piwetz concludes that repeated correspondence between direction-finding and visual observations identified the light as the signal source. That conclusion belongs to the wing-intelligence author. Its supporting chronology also records differences between the channels, notably the 10.57 electronic observation without ground contact. Correlation in a written account is consequential evidence, but its timing precision cannot be independently tested from narrative timestamps alone.
The Blue Book card identifies American Airlines Flight 655. The file's 8 November CAA letter instead discusses Flight 655 encountering Flight 966 near El Paso, with 966 departing for Dallas. It does not identify the RB-47's entire Mississippi–Texas–Oklahoma sequence. The card's aircraft classification is an official recorded disposition whose documentary basis needs scrutiny, rather than a complete demonstrated explanation of the crew's observations.
The Colorado project's Case 5 depended on interviews roughly ten years later and reported failing to recover the original 1957 file. Its conclusion therefore described inadequate recollections for identification. Recovery of the 1957 documents changes that evidential premise. Gordon Thayer's separate chapter considered optical mirages and anomalous radar propagation, but used weather profiles for 18–19 September 1957, rather than 17 July. He also judged the combined proposed solution improbable and retained an unknown classification. Those profiles cannot establish atmospheric conditions on the actual flight date.
McDonald's 1971 article contrasts interviews with the recovered documents and emphasises apparent simultaneous loss and return across visual, electronic and ground channels. It acknowledges imperfect recollection and timing discrepancies. The AIAA subcommittee published the case for readers' assessment and reported Chase's approval; that editorial setting is not an independent instrument validation or a formal institutional conclusion about the source's origin. Its reproduced article also says forwarding occurred on 17 November, whereas the actual endorsement plainly reads 17 October.
Klass proposes a specific combination: a temporary direction-finder fault reversing the initial bearing, later reception of Biloxi/Duncanville ground radar, a meteor for the first brief visual crossing, and ordinary lights or aircraft for subsequent observations. His circuit argument identifies a possible relay mechanism for a 180-degree error; no inspected maintenance record demonstrates that fault on this aircraft. His bearing reconstruction depends on inferred aircraft position and heading, and he acknowledges imperfect agreement. This is a testable engineering reconstruction, rather than a documented repair diagnosis.
Klass suggests Flight 966 may account for a later Dallas-area radar/light observation, using its scheduled arrival. His paper explicitly calls the identification speculative and says original flight records could not then be located. The CAA letter places a Flight 966 encounter near El Paso, but supplies no arrival log or synchronised Dallas track. The inspected material therefore does not resolve whether that aircraft explains the relevant later phase.
The competing reconstructions turn on source association. McDonald gives greater weight to the crew's reported correlation; Klass separates the episode into different lights and ground transmissions. The surviving material establishes an extensively reported operational experience with specialised observers and recoverable documents from 1957. It leaves the complete correspondence between the signals, lights and ground returns open to further technical reconstruction, rather than establishing an extraordinary emitting craft from the number of reported channels alone.
Key dates
17 July 1957
Crew reports unusual electronic indications and intermittent lights during Lacy 17's southern-US flight; Duncanville sends a same-day report.
15 August 1957
Air Defense Command asks for the Airborne Observer's Data Sheet from Major Chase.
10 September 1957
Chase completes the questionnaire, describing light without determined dimensions and distinguishing failed airborne-radar detection from reported Utah tracking.
17 October 1957
The dated forwarding endorsement sends the documents for analysis and records the command's unknown assessment.
8 November 1957
CAA correspondence identifies another airliner in the separate Flight 655 encounter near El Paso.
1967–68
Colorado investigators interview crew and search unsuccessfully for the 1957 records under an incorrect incident date.
1969
McDonald conducts the interviews used alongside the recovered file in his later article.
July 1971
AIAA's magazine publishes McDonald's sample-case reconstruction and distinguishes it from the earlier date-limited Colorado analysis.
30 December 1971
Klass's paper concludes with a conventional reconstruction and selected favourable reactions attributed to Chase and McClure.
57-page mixed compilation, containing both 1957 records and later additions. Selected scans inspected; page 1 record card, pages 3–4 request/endorsement and pages 5–6 assessment notes used. Page 4 visibly 17 October 1957. The complete compilation was not reviewed page by page.
All twelve questionnaire pages inspected as scans, with non-sequential compilation order. PDF41: failed airborne attempt, reported Utah tracking and scope-picture note. PDF48: only light/no dimensions. PDF52: role and completion date. No original scope pictures inspected.
PDF37–40 fully read as scans. Elwin T. Piwetz identified as wing intelligence officer. Precise report-writing day not established. Source's correlation conclusion attributed; 11.02 timestamp retained as printed, with later 11.20 transcription distinguished.
Three split message pages read at PDF54, 13 and 35 (also reproduced partly on 24/46). Contains aircraft-reported contact and ground-contact limitation; handwritten annotations have unidentified authors and dates and are not adopted as original message content.
All five re-typeset PDF pages read and compared with selected magazine scans in the official compilation. Includes AIAA introduction, map and source-owned reconstruction. Primary October forwarding date overrides its November statement. Original interview tapes and Chase approval letter not inspected.
Full short case text read through NCAS transcription. Retrospective interviews and unsuccessful record search, including recording-history contradictions; predates recovery of the 1957 file. No original interview recordings examined.
Relevant Case 5 section read, NCAS lines 426–477. September weather profiles expressly dated differently from actual July event. Thayer's tentative unknown conclusion and proposed explanations retained with that date limitation.
Selected original paper pages read as scans: PDF15,16,20,23,25,29–31; supporting extracted passages also read. Printed16/PDF31 dated 30 December 1971 includes selected Chase/McClure responses. Original correspondence, complete equipment manuals, maintenance diagnosis and airline arrival records not independently inspected.
Entire PDF57 scan read. Distinguishes Amarillo TWA episode from Flight655/966 encounter near El Paso. Does not supply a full RB-47 route identification or actual Dallas arrival log.