Kevin Randle published a blunt conclusion on his blog in late January 2026, ending a weeks-long argument with a French Roswell researcher who goes by Gilles: the debate, he wrote, was a waste of time with someone who simply could not understand the facts. The dismissal was pointed enough to read as finality. But the argument itself — about a balloon flight, a missing diary entry, and a chain of unrecorded interviews — exposes something worth examining carefully, because the evidentiary problems Randle identified don't all belong to the other side.
The story being told here is narrow but significant. The primary prosaic explanation for the Roswell debris — that it came from Project Mogul, a classified Cold War program using balloon arrays to monitor Soviet nuclear tests — depends substantially on a flight known as Flight #4. And Flight #4, by the documentary record Randle himself cites, was cancelled.
What the Documentation Actually Shows
Randle, a prolific Roswell investigator and author, has spent more than three decades reviewing the project logs from the New York University balloon program that supplied the hardware for Mogul. His reading of those records is unambiguous: Flight #5 was the first successful flight in New Mexico. The NYU documentation, he writes, records Flight #5 as carrying a full load with telemetry — and a B-17 and personnel were dispatched to recover it some 25 miles east of Roswell. Flight #4, by contrast, does not appear in the record as a completed mission.
The complication is Charles Moore. Moore, one of the NYU scientists who worked on the balloon program, told Randle — without, Randle emphasizes, any supporting documentation — that Flight #4 had in fact been launched as a full array, around three in the morning, in violation of the operating rules. Moore further claimed that Flight #4 was configured the same way as Flight #5 and performed as well, if not better. Randle's response to that claim is pointed: if Flight #4 was cancelled, as the written record indicates, it couldn't have generated data — and Moore's assertion that it performed well is therefore impossible to square with a cancellation. "If it was cancelled," Randle writes, "then it couldn't have dropped debris on the ranch."
"then it couldn't have dropped debris on the ranch."
Gilles, pressing from the other direction, asked Randle repeatedly what time Flight #4 was launched — because Moore, across various interviews and accounts, has provided more than one answer. Randle acknowledges Moore supplied multiple conflicting times. That matters, because researchers attempting to calculate whether a flight launched from Alamogordo could have deposited debris on the Brazel ranch north of Roswell depend on launch time, wind data, and trajectory modeling. If the launch time itself is uncertain, those calculations carry significant error bars. Randle concedes the preponderance of available evidence suggests Flight #4 would not have come within 17 miles of the actual debris field — and notes that 17 miles is itself a substantial distance from where Mack Brazel found the material.
What documentation exists to support Moore's claim that Flight #4 flew as a full array? That question remains open. None is cited in any of the exchanges under review.
The Witness Problem Runs Both Ways
The Flight #4 dispute is, at its core, an argument about whose unverified testimony to credit. And here is where the exchanges between Randle and Gilles reveal something the Roswell debate rarely surfaces cleanly: the evidentiary standards applied to witnesses are not consistent, and both sides have benefited from the inconsistency.
Randle confronts this directly. Gilles had cited statements attributed to Brigadier General Thomas DuBose — statements that, in Gilles's framing, supported the idea that the material photographed in General Roger Ramey's Fort Worth office was the actual Roswell debris, not a substituted weather balloon. Those statements came through Jaime Shandera, a researcher who interviewed DuBose but, by Randle's account, neither recorded the conversation nor took notes. Randle writes that Shandera did not record his interview with DuBose, nor did he take notes.
Randle had separately located and interviewed J. Bond Johnson, the photographer who took six pictures in Ramey's office in July 1947. By Randle's account, Johnson told him on tape — and repeated it six or seven times — that the material in those photographs was a weather balloon. After Shandera and Bill Moore subsequently interviewed Johnson, Randle writes, Johnson's account shifted: the debris in the pictures became the real material, which in turn created a need for DuBose to say no substitution had occurred. Randle holds a taped affidavit from DuBose that, he says, confirms the material was switched — directly contradicting what Shandera reported DuBose had said.
The question Randle poses to Gilles is precise: why accept Shandera's unrecorded account of what DuBose said, while rejecting Randle's unrecorded account of what Johnny Mann — the filmmaker who conducted early Marcel interviews — told him? Mann, according to Randle, saved the raw footage of Jesse Marcel's interview and later conceded that reusing the tape for what he then considered a more important story, the Pascagoula abduction, had been a mistake. Gilles dismissed Mann's account partly on the grounds that it was unrecorded. Randle's point is that by that standard, Shandera's DuBose account should be dismissed too.
This is not a trivial methodological dispute. It goes to the heart of how the Roswell case has been built — and how it has been argued against.
A Newspaper Article and a Second-Hand Chain
In late February 2026, Randle highlighted a finding by researcher David Rudiak: a July 9, 1947 edition of the Dayton Herald, reporting that Wright Field would not receive the weather device previously believed to be a flying disc. The paper quoted a message from Army Air Forces headquarters to the office of technical intelligence: "In view of positive identification of this object as a Rawin high altitude sounding device (radar target) it appears unnecessary to forward it to you."
This matters because the FBI teletype sent from Fort Worth on the evening of July 8 — based on an interview with an officer identified in the document as Major Kirton — stated that the disc and balloon were being transported to Wright Field by special plane for examination. The Dayton Herald article, if accurate, indicates that plan was reversed after the Fort Worth identification was made.
Randle interprets this as evidence that General Ramey's weather balloon identification was genuine and complete, making a Wright Field shipment unnecessary. But he also acknowledges a counterreading: Albert Trakowski, a Mogul project officer, told Air Force investigators in 1994 that Colonel Marcellus Duffy had called him from Wright Field to say that someone had brought him debris from New Mexico that looked like material launched from Alamogordo. Athelstan F. Spilhaus, another NYU figure, wrote that the debris may have been sent to Duffy specifically because he was associated with the project.
These are second-hand accounts, gathered nearly fifty years after the fact, about what Duffy may have done. Duffy's own words, if he left any record, are not part of the available documentation. Trakowski said Duffy was not concerned about a security breach. But whether the debris reached Wright Field at all — in any quantity, in any form — is a question the newspaper article raises without settling. Randle himself puts it plainly: this boils down to which witnesses you choose to believe.
Concerning a cover story for the project, Trakowski told Air Force interviewers that there was no planned cover story and that he could recall no documentation or efforts to develop one, even though security for Mogul was of great concern. Randle finds that plausible, noting that what the NYU team was actually doing in New Mexico was not classified — balloon experiment photographs appeared on newspaper front pages, including a July 10 press event in Alamogordo. The ultimate purpose of the program, monitoring Soviet nuclear activity, was classified. The hardware was not.
And yet, if the hardware was unclassified and the Fort Worth identification was genuine, why did Marcel — the base intelligence officer who had handled the debris — later say the material came from something not made on Earth? Randle does not dismiss Marcel. He notes that Marcel had not seen the bodies allegedly recovered at a second site, knew about them only through rumor, and declined to discuss what he hadn't witnessed. What Marcel had seen, he stood behind.
Two sites, Randle argues, eliminate Mogul as an explanation regardless of what Flight #4 did or didn't do. A balloon array, cancelled or not, does not produce a second debris field containing what multiple witnesses described as recovered craft and bodies.
The specific document worth watching is any contemporaneous record — diary entry, flight log, or internal NYU communication — that addresses Flight #4's status on June 4, 1947. Charles Moore's account is the architecture on which the Mogul-Roswell connection rests. If the paper trail for that single flight surfaces, or is conclusively shown not to exist, the argument changes shape entirely.