On the night of 5–6 September 1966, reports of a changing-colour light in northeastern Minnesota prompted attention at Finland Air Force Station and an attempted interception by aircraft from Duluth. The surviving file contains several perspectives: a civilian call, airmen’s visual observations, radar reports, a follow-up investigation and an engineering assessment. They do not form an entirely consistent account of one tracked object.
Finland, Minnesota, rather than Finley
The eleven-page Project Blue Book packet concerns Finland Air Force Station in Minnesota. A covering letter of 27 September explicitly corrects the earlier reference to Finley Air Force Station. That correction matters when tracing the case: an initial message can carry a place-name error forward even when a later document resolves it. The associated civilian reports came from the North Shore area near Lutsen and Grand Marais. They should not be reduced to a single exact observation point.
The packet is also out of chronological order. The four-page incoming message is distributed across the beginning and end of the PDF, with subsequent correspondence in between. Reading only its first page would miss the limitations expressed in the later assessment, while reading only the assessment would omit what the observers actually reported.
A civilian call and the station’s response
In his report dated 19 September, Second Lieutenant Theron J. Sumner wrote that the station became involved after a civilian reported an unusual object. The call was relayed to the air-defence organisation at Duluth. Airmen then looked for something corresponding to the report. Sumner recorded a claimed green light near the caller’s home, but that account does not independently establish the light’s distance or altitude. The private residence is not identified on this page.
One airman’s statement, preserved separately, describes a mostly white object with a red fringe and an oval appearance. He said it seemed to move while remaining in the same general area. He then went to the station’s equipment and reported a radar return. This is an account of how the observer associated the visual experience with a display indication; it does not itself demonstrate that the light and the radar return shared a physical source.
What the radar operators described
Sumner reported that the station initially obtained no return in the relevant area and that clutter made observation difficult. Later, technicians described a return on the FPS-90 display that changed range between approximately 13 and 54 miles while remaining along the same westerly bearing, 270 degrees. Individual appearances lasted a few minutes before disappearing and recurring at another range.
Two F-89 aircraft searched at reported altitudes of about 8,000–10,000 feet. Here the documents diverge. The initial message says the aircraft converged with the radar indication but obtained no airborne sighting, with an apparent altitude mismatch. The later account says the return moved away as the interceptors approached. Neither version supplies a successful close visual identification of an unusual aircraft.
These differences are material to the interpretation. Convergence without a sighting and a return that changes position before an aircraft reaches it are different observational circumstances. Combining them into a seamless story of an object repeatedly evading fighters would remove the very uncertainty that the contemporary reviewers discussed.
The following night and the missing photographs
Sumner also observed the changing-colour light when it reappeared the next evening. He judged it to be a star low near the horizon, which would subsequently set. He noted that stars visible from the mountain could also sparkle in several colours. This is a specific follow-up observation by the investigating officer. It is stronger than an explanation inferred merely from a later catalogue label, while still not being a direct measurement of every light described by every observer.
The packet records attempted photography, but the follow-up says no photographs were produced. Incorrect camera settings and unsuitable film were given as reasons. An early message’s suggestion that personnel had photographed the object therefore cannot be treated as an available evidential photograph. The image accompanying this article is the official written assessment, not a recovered image of the reported phenomenon.
The 27 September covering letter adds reports from the following evening and says they might be related. It also describes the accounts as confused and contradictory and proposes stars, temperature inversions, ground clutter or anomalous radar propagation as possibilities. The existence of another report is preserved; a common cause is not assumed simply because the reports occurred nearby.
The October engineering assessment
On 12 October, a Foreign Technology Division memorandum signed by Colonel Jerome J. Jones endorsed the concern about inconsistent reporting. It explicitly compared the initial convergence account with the later account of the return moving away. The reviewer also considered the lack of a visual target despite the reported visibility and interception geometry.
The memorandum proposed that atmospheric conditions could produce erratic radar indications. A return that changed range while retaining one bearing suggested reflections associated with a fixed ground feature rather than demonstrating an object travelling along the displayed track. Crucially, the assessment said that incomplete details prevented a more definite evaluation of the radar portion. It offered an argued explanation with a stated limit.
The later Colorado analysis
In the Colorado project’s optical and radar chapter, case 1321-B, Gordon O. Thayer favoured a star combined with anomalous radar propagation. He proposed Shaula as the likely star and compared the reported radar heights with atmospheric-layer calculations using available weather observations. That modelling depended on whether the relevant layer existed over Finland at the assumed elevation. Thayer also acknowledged uncertainty over why the returns were confined to such a narrow direction. His analysis supports a possible natural account; it is not a direct contemporaneous measurement of the target or of every atmospheric condition along the radar path.
Location and outcome
The pin is a broad Finland locality reference from the US Census Bureau’s Minnesota places gazetteer, rounded to three decimal places. It is not the exact radar site, a private witness’s home, a fighter-interception position or a position of the reported light. This historical record does not establish public access to former military property.
Outcome: Possibly Explained. The officer’s follow-up star identification and the qualified atmospheric radar assessments support a natural explanation for substantial parts of the report. The contradictory intercept accounts, incomplete details and absence of produced photographs prevent a claim that every observation has been demonstrated to share that explanation. The record card itself preserves insufficient data for evaluation alongside its radar propagation assessment.
Source review: 13 November 2026. All eleven pages of the released file were read, with the record card, Sumner report and October assessment checked against their page images. The relevant Colorado analysis was read as a later interpretation. No independent radar reconstruction, astronomical calculation or reanalysis of original photographs is claimed.