The Department of War published its fourth PURSUE tranche on Friday, July 10, 2026 — four weeks after the June 12 release, holding the Friday cadence the program has kept since it began. The batch comprises 40 records, the smallest yet, and it moves the program’s center of gravity backwards in time. Where Release 01 was archival breadth, Release 02 was operational sensor video, and Release 03 belonged to the FBI, Release 04 belongs to the founding era: the paperwork of the first American attempts to decide what, if anything, was in the sky. Every record is indexed in the full PURSUE archive, and the running corpus now stands at 334 released records across four tranches.
The document at the center of the tranche is the Project Sign progress report of 1948 — the initial report of the Air Materiel Command program that began the U.S. government’s investigation of unidentified flying objects and that gave way, in turn, to Grudge and then Blue Book. It catalogues 100 sightings from 1947 and 1948, and it arrives bound together with an article excerpted from the British aviation periodical The Aeroplane under the title “The Biology of the Flying Saucer.” Alongside it sit two Air Intelligence Division studies of flying object incidents, from 1948 and 1949. The second, Study No. 203, is the more quotable and the more honest: it assesses that “it appears that some object has been seen; however, the identification of that object cannot be readily accomplished,” offers two origins it considers reasonable — technology of domestic or of foreign make — and recommends that, if foreign, the United States assume Soviet scientific or military activity and take the threat seriously. That is the whole of the early official position in one sentence: something is up there, we cannot name it, and the candidate we fear is Russian.
The end of that era is in the tranche too. The Department of the Air Force ad hoc committee convened in 1966 and 1967 to review Project Blue Book recommended that the Air Force contract a university-affiliated scientific team to investigate selected sightings. The Air Force adopted the recommendation, and what followed was the Condon Committee, whose 1968 report concluded that further study was unlikely to be scientifically justified and gave the Air Force the grounds it needed to close Blue Book in 1969. The file is, in effect, the paper trail of the decision to stop looking. Around it the release stacks 220 pages of Blue Book correspondence from 1955 — letters between government departments, Congress, and members of the public — which is less a case file than a portrait of how much of the program’s energy went into answering the mail.
The most quietly deflationary document in the release is the file on joint U.S.–Canadian aviation projects. It concerns the Avro “Project Y2,” a near-circular vertical take-off aircraft under joint development, and it contains a 1954 memorandum observing that a circular-planform VTOL craft may well be mistaken for a UFO by observers unfamiliar with the technology — and recommending that UFO reports near Soviet assets be re-examined to see whether they might be advanced foreign aircraft of exactly this kind. The United States government was, in other words, building a flying saucer, and writing memos to itself about how its own saucer would be misreported. Any serious accounting of the 1950s sighting wave has to hold that document in one hand.
It has a companion. The CIA’s contribution is a 1955 memorandum debriefing four men who watched a luminescent greenish-yellow object from a train crossing Soviet Azerbaijan between Baku and Tiflis. The June tranche released the sighting report; this one names the party, and the name matters — among the observers was U.S. Senator Richard Russell, then one of the most powerful men in American defense politics, travelling with a military officer and two government officials. It is as close to an unimpeachable witness roster as the mid-century record offers. The Agency’s own conclusion, in the same document, is that the sighting can “probably be explained as steep climbing aircraft or missiles,” and that “the evidence does not appear sufficiently firm to warrant the conclusion that the Soviets have developed a radically new type of aircraft.” A contemporaneous analysis released with it reaches the same place and cites Robertson Panel findings that almost none of the sightings represented a threat. This is the pattern the historical half of PURSUE keeps producing: the witnesses get more credible and the conclusions get more mundane, and the two facts are not in tension. A senator can see something and be wrong about what it was.
At the modern end, the tranche does something the archive has not seen before. Three U.S. Navy Range Fouler debriefs — the standardized form used to record an unauthorized intrusion into controlled airspace — were released alongside the sensor videos they describe. In the 2020 Eastern United States case, an operator’s written account of an object that was “quite small,” of “indistinguishable” shape, metallic, with a reflective underside, is published together with the footage they were watching. Every other video in the corpus arrives as an unaccompanied clip, and every other testimony arrives without an instrument. Three times in this release, they arrive together. That is the single most evidentially interesting structural fact in Release 04, and it deserves to be said plainly: it does not make the objects anomalous, but it does make the claims checkable, which is a different and rarer thing.
The other nineteen videos are the familiar material — short infrared clips from INDOPACOM over the Yellow Sea and the East and South China Seas, from NORTHCOM over the eastern and western United States, from CENTCOM over the Middle East, spanning 2015 to 2025, each carrying AARO’s standing caution that the shot-by-shot description should not be read as an analytical judgment. One is worth singling out for its candor. The 1996 Western United States video is accompanied by AARO’s admission that no formal data-handling practices for UAP records existed when it was reported, that the media was digitally altered before it reached the task force, and that it is presented as received. The clip itself is three seconds of an object crossing the frame, then replayed, then slowed, then frozen. A release program willing to publish that sentence about its own evidence is behaving better than one that isn’t — and the clip remains worthless as proof of anything.
NASA’s four audio records are the tranche’s most useful corrective. The Apollo 14 and Apollo 17 post-mission crew debriefings capture astronauts and flight surgeons working through the “light flash phenomena” — streaks and flashes the crews saw, sometimes with their eyes closed, in lunar orbit and on the surface. The government’s own summary supplies the answer: it is a now well-documented biological effect in which high-energy cosmic rays pass through the eye and strike the retina. These recordings are in a UAP file, and they are not a mystery. They are a question being resolved in real time, and their presence is a standing argument against the assumption that everything the government filed under “unidentified” stayed that way. The three STS-80 photographs, taken from Columbia in low-Earth orbit in late 1996, are the opposite case and should be treated with the caution the orbital environment demands: ice crystals, debris, and thruster effluent are abundant in that frame and photograph beautifully.
The most revealing item in Release 04 is administrative. The 2015 Pantex Plant incident report documents an object intruding on the airspace above the Amarillo facility where American nuclear weapons are assembled and taken apart — and pages five and six of it, the Department of Energy notes, were already released under PURSUE on May 22, in a more heavily redacted form, as the Enhanced Pantex Imagery file. Seven weeks later the same pages are back with less blacked out. Nothing in the program’s public framing prepares you for that, and it is the first hard evidence that PURSUE’s redaction decisions are being revisited rather than simply executed once. It also quietly extends the pattern the corpus has traced across the nuclear complex since 1948.
What Release 04 establishes is not a case but a shape. Read end to end, the founding-era files show an investigative apparatus that suspected the Soviets, knew its own experimental aircraft were being mistaken for saucers, could not get its own physicists to agree in a room at Los Alamos, and eventually commissioned the study that let it stop. None of that resolves a single sighting, and none of it is an argument that nothing is there. It is an argument that the government’s uncertainty was real, documented, and, by its own account, never dispelled — which is a smaller claim than disclosure’s loudest advocates make and a larger one than its critics allow. The program overview is at /pursue/, and every document page is indexed at /disclosure/archive/.