← Back to Files & Wiki
Wiki page · asset · graph_investigation_draft

65_HS1-834228961_62-HQ-83894_Section_3

Official PDF: Open Sky release file copy · WAR.GOV source

Release 01#war-gov#pursue#release-01#official-source#evidence#graph-investigation-draft#pdf#fbi

65_HS1-834228961_62-HQ-83894_Section_3

Evidence media

Selected page renders below are derived from the official PDF to show the packet's evidence types without turning the 190-page file into a gallery.

Derived page render from official PDF: confidential debris enclosure photograph Page 22: confidential enclosure photograph showing an irregular dark debris or fragment-like object on a marked page. Treat it as an exhibit photograph, not as aerial-object imagery.

Derived page render from official PDF: Twin Falls newspaper artist's conception Page 128: Twin Falls newspaper clipping with an artist's conception of a reported flying disc, including side, bottom, and end-view drawings. This is a press depiction preserved in the file, not a photograph of the reported object.

Derived page render from official PDF: A. C. Urie handwritten sketch page Page 133: handwritten A. C. Urie sketch/statement page with a side-view drawing and notes about a reported Snake River canyon object. It should be handled as a witness drawing and statement exhibit.

Derived page render from official PDF: alleged specimen photograph Page 167: exhibit photograph labeled "FLYING SAUCER" ALLEGED SPECIMENS, showing assorted fragments and components laid out on a surface. The adjacent text pages are needed for custody and identification context; the image alone does not establish origin.

Investigation reading

This is a 190-page FBI Headquarters section from file 62-HQ-83894, preserved in WAR.GOV/PURSUE Release 01 as an official PDF asset. It should be read as a packet of early flying-disc serials, transmittals, witness statements, photographs, newspaper clippings, and interagency coordination records, not as one single incident.

The packet is heavily centered on the 1947 wave. The recurring anchors are Kenneth Arnold, Harold A. Dahl, Fred L. Crisman, Emil J. Smith, Ray Palmer / Venture Press, Fourth Air Force intelligence at Hamilton Field, Air Defense Command, and FBI field offices. It also contains separate report threads from Muroc Army Air Field, Twin Falls / Snake River canyon, West Rindge, Guam, Placerville, Montana, Oklahoma City, Philadelphia, Utah, Seattle/Boeing, and other places. Several pages are reverse sides, routing sheets, envelopes, or mostly blank archival pages.

The strongest reading is evidentiary and custodial: Section 3 shows how the FBI and Army Air Forces were exchanging, triaging, and sometimes disputing responsibility for flying-disc reports in mid-to-late 1947. Some pages preserve witness accounts; some preserve agency skepticism or prosaic leads; some preserve photographs of alleged debris/specimens. This draft does not resolve any report, endorse any claim, or classify any case as explained or unexplained.

Full-source pass completed for this draft: the released PDF has 190 pages, file size 50,907,255 bytes, and SHA-256 1a9e21457d266c408a9982485034fed73ebdbede94b3c8322ee024eaa60cfe57. The OCR text has 190 page markers, 189 pages with text, about 296k OCR characters, and 160 OCR chunks. Low-text or mostly non-substantive pages include page 7, 57, 62, 96, 126, and 190. Representative rendered-page checks were made for the debris/photo pages, Twin Falls newspaper/sketch pages, the alleged-specimen memo/photo pages, and the closing memorandum that references attached photographs.

What the file appears to contain

At packet level, Section 3 contains these major page ranges:

PDF pagesPacket content
1FBI cover / declassification and records-center markings for Section 3.
2-4San Francisco transmittal and Fourth Air Force correspondence about F. M. Johnson of Portland, who wrote that he saw similar objects from the Mount Adams district on June 24, 1947, around the Kenneth Arnold sighting period. Page 3 notes the Fourth Air Force's concern that Johnson may have read about Arnold's report in newspapers and asks for any FBI interview result.
5-9Transmittals from Lt. Col. Donald L. Springer. Page 6 preserves a Weekly Intelligence Summary extract about three enlisted men at Harmon Field, Guam reporting two small crescent-shaped objects on August 14, 1947.
10-13Air Rescue Service Final Mission Report for Mission Number Thirty-Nine, the August 1, 1947 B-25 crash near Kelso, Washington. These pages describe the crash response, survivors, fatalities, classified material concerns, ground and aerial activity, and photographs as enclosures.
14-19Fourth Air Force CI-RI report on the Tacoma / Maury Island flying-disc investigation, including Arnold, Capt. Davidson, Lt. Brown, Emil Smith, Fred Crisman, Harold Dahl, Ray Palmer, and Venture Press. Pages 18-19 include Venture Press correspondence and a telegram noting Palmer sent Arnold funds to investigate the Tacoma report.
20-25Official photograph/enclosure pages. Rendered-page inspection confirms pages 22-25 are photographs of dark irregular debris or fragment-like material, not narrative text pages. The images are marked as confidential enclosures and should be treated as exhibit photographs rather than aerial-object imagery.
26-28Boston / West Rindge, New Hampshire memorandum about metal fragments observed July 7, 1947. The memo says spectrographic examination found ordinary cast iron subjected to high heat and says no further Boston action was planned unless requested.
29-30Twin Falls, Idaho teletype about A. C. Urie and his sons reporting an object near the Snake River area, with speed, size, color, flame, and sound claims recorded from the newspaper/witness chain.
31-59Additional FBI/Seattle/Tacoma material on the Dahl-Crisman-Maury Island story, anonymous press calls, Arnold and Smith interviews, alleged fragments, the B-25 crash context, and related statements/depositions. The packet records both extraordinary claims and strong internal skepticism; those are source positions, not findings by this page.
60-72Muroc Army Air Field investigation packet. It includes statements attributed to Colonel Gilkey, Major Richard E. Shoop, T/Sgt. Joseph Ruvolo, Lt. Joseph C. McHenry, Major J. C. Wise, Captain John Paul Stapp, Jannette Marie Scott, and S/Sgt. Gerald W. Bauman. The pages describe metallic/spherical/disc-like objects, weather-balloon comparisons, altitude/speed estimates, and aircraft-test context.
73-84Sacramento / Placerville and Montana-related report material, including Ray A. Switzer's August 14, 1947 observation near Placerville, the E. J. / R. J. Madden Canyon Ferry report, and Ward L. Stewart's report after a training flight near Berkeley.
85-88Twin Falls Hawkins/Brown report and a San Francisco memorandum seeking interview/photos from William Rhoads of Phoenix, Arizona, who was reported to have photographed objects on July 7, 1947.
89-95Butte / Boise interview material on Kenneth Arnold and the Tacoma/Maury Island story, including Arnold's account of contacts with Dahl, Crisman, Smith, press calls, and Army officers.
97-114Venture Press letters, Arnold article material, Arnold biographical pages, and postal/registered-mail records. Pages 105-111 preserve Arnold's first-person account of the June 24, 1947 Cascade/Mount Rainier sighting as prepared for publication or official submission.
115-123Teletypes for Seattle/Boeing reports, Oregon/Agate Beach, Philadelphia, and Twin Falls routing; page 123 introduces the newspaper account and Urie's own sketch sheet.
128-134Twin Falls newspaper clipping and A. C. Urie drawings. Rendered-page inspection confirms page 128 contains a newspaper clipping with an artist's flying-disc diagram labeled side, bottom, and end views with rough dimensions; pages 133-134 contain hand-drawn side/top/bottom sketches and notes about Snake River canyon. These are drawings/clippings, not photographs of an object.
135-142Memos concerning Byron H. Savage of Oklahoma City and Kenneth Arnold background material. Savage's report includes a round/elliptical object, a possible swishing sound, and uncertainty about the sound.
143-156Additional teletypes and field memoranda, including reports attributed to Brummett, Richard F. Shaver background checks, Anchorage / Fort Richardson officers, and the Philadelphia August 6, 1947 object report.
157-161FBI / Army Air Forces policy exchange about who should investigate flying-disc reports and ground objects. These pages are important for institutional context because they show tension over scope and workload.
162-168Danforth, Illinois / alleged-specimen thread. The packet discusses an instrument said by one contact to relate to classified Air Force testing and a Wright Field memo on alleged "Flying Saucer" specimens. Rendered inspection shows pages 164-165 are text pages describing ordinary or identifiable components and saying the specimens had no connection with MOGUL or any AMC project; page 167 is a photograph of assorted alleged specimens; page 168 appears to be the blank/back side of an official photo sheet with Wright Field/confidential markings.
169-173Salt Lake / Loren, Utah reports and another copy of the Madden / Canyon Ferry report. The Salt Lake memo records multiple couples reporting moving lights/objects around September 8, 1947, while noting the information was indefinite.
174-177Post Office / Bogue Foundation correspondence about sale of information on flying discs.
178-180Bruce Armstrong / M. A. Nichols Seattle/Boeing report material. Page 178 summarizes a Seattle Post-Intelligencer article about a dark or blue triangular object near Boeing; page 180 records a prosaic lead involving burnt paper blown over the field for at least one reported object, while also preserving Nichols's separate high-altitude report.
181-185San Francisco transmittal of Air Defense Command cooperation guidance, plus Bureau Bulletin language saying FBI investigative activity under the earlier flying-disc bulletin was discontinued and future reports should be referred to the Air Forces.
186-188Butte memorandum about a May 1947 silver-object / gaseous-pillar report between Ellensburg and Seattle, plus an October 9, 1947 memorandum stating that photographs of various Air Corps reports were attached for Bureau files. Rendered inspection of page 188 shows a memo referencing photographs, not the photographs themselves on that page.
189-190Late reverse/marking pages with little substantive text.

Several source threads recur across different serials. The Maury Island / Tacoma material appears in early Fourth Air Force narrative pages, later FBI memoranda, Arnold/Smith accounts, alleged-fragment photos, and press-call records. The Twin Falls material appears first as a teletype, later as a memorandum, newspaper clipping, and witness sketch pages. Arnold material appears both as biographical background and as a first-person sighting article. These repetitions are useful for provenance, but they can also inflate graph claim counts if treated as independent events.

The visual pages need special handling. Pages 22-25 and 167 are photographs of physical debris/specimens. Page 128 is a newspaper reproduction with an artist's concept drawing. Pages 133-134 are Urie sketch/diagram pages. The packet contains many references to photographs, negatives, and cameras, but not every reference is accompanied by a visible image in this PDF.

Source custody and provenance

The release manifest description says this file is part of the FBI 62-HQ-83894 case file and that the FBI Vault version is partial, with more redactions and some pages missing. That statement is provenance context from the manifest and still needs page-by-page comparison before using it as a detailed declassification claim.

Graph context

Open Sky's graph currently models this released asset as an official Document node and also has a related release-record Document for the WAR.GOV CSV row. The document node preserves the official URL, canonical URL, release dataset, SHA-256, and official-primary provenance tier.

The semantic integration layer currently has:

  • 1,159 extracted claim records,
  • 512 entity-mention records,
  • 42 sensor-event records,
  • 0 extracted table rows,
  • 160 OCR chunks for this asset, plus release/manifest text chunks around the document record.

Those graph records are investigation scaffolding. They should not be read as human findings. In this packet, many extracted "sensor" records are actually ordinary mentions of aircraft, camera, photograph, control-tower, or witness-observation text inside historical documents. The graph also has 10 CANDIDATE_CROSSLINK leads. Several are triggered by the word "Minot" in Kenneth Arnold's biographical pages, not by a Minot incident in this Section 3 packet. Those links need human review before they are promoted as real case connections.

Useful graph interpretation for this page:

  • Keep Section 3 as a parent packet node until child case pages exist.
  • Treat repeated serial copies and newspaper reproductions as separate provenance records for the same underlying threads, not automatically as independent corroboration.
  • Treat the debris/specimen photographs as exhibit records requiring visual/source custody review, not as aerial-object photographs.
  • Keep Maury Island, Muroc, Twin Falls, Arnold, Philadelphia, Utah, Seattle/Boeing, and policy-discontinuance material separated into child clusters before cross-case analytics.

Leads to check

  • Split this Section 3 packet into child pages or case clusters for: F. M. Johnson / Mount Adams; Guam / Harmon Field; Kelso B-25 crash; Maury Island / Tacoma / Dahl-Crisman-Arnold-Smith; West Rindge fragments; Muroc AAF statements; Twin Falls / Snake River / A. C. Urie; Phoenix/Rhoads photographs; Byron Savage / Oklahoma City; Philadelphia August 6 report; Utah / Loren reports; Seattle/Boeing Armstrong/Nichols; Danforth/Wright Field alleged specimens; and FBI/AAF policy coordination.
  • Compare this WAR.GOV PDF against the available FBI Vault section page by page. The manifest says the Vault version is more redacted and incomplete; that needs a concrete diff, not a general statement.
  • For pages 22-25 and 167, locate any original photograph logs, enclosure indexes, or lab records that identify who collected the fragments/specimens and whether the exhibit photos correspond to Maury Island, Danforth, Wright Field, or another thread.
  • For the Maury Island/Tacoma material, build a provenance timeline that separates Dahl/Crisman statements, Arnold/Smith statements, press calls, Palmer/Venture Press correspondence, B-25 crash records, and FBI/Army narrative summaries.
  • For Muroc pages 60-72, extract witness statements into a structured child cluster and run prosaic checks before any escalation: aircraft activity, test schedules, balloon releases, sun angle/reflection, weather, and distance-estimate uncertainty.
  • For Twin Falls pages 123 and 128-134, separate the newspaper artist's conception from Urie's own sketches and from the teletype summary. The drawing is evidence of reporting/depiction, not a direct image of the object.
  • For the Phoenix/Rhoads note on page 87, search for the actual photographs/negatives and any interview records before treating the photo reference as represented in this PDF.
  • For the Danforth/Wright Field specimen thread, reconcile page 162's instrument/MOGUL discussion with pages 164-168's separate specimen memo/photo. The visible Wright Field memo says the alleged specimens had no connection with MOGUL or AMC projects; do not merge that with all Danforth instrument references without a chain-of-custody check.
  • Review OCR quality on low-text/reverse pages and any photograph pages before using absence of text as absence of content.
  • Review the current graph crosslinks around "Minot" so Kenneth Arnold biographical references do not become false incident correlations.

Lead check notes

  • Partial — child-page split: Current linked Release 01 corpus hits support splitting this packet into child clusters rather than treating Section 3 as one case. Maury Island/Tacoma appears across Sections 1, 2, 3, and 7; Muroc appears in Section 3 and the incident-summary file; Twin Falls/Urie appears in Section 3, Section 2, Serial 130, and the incident-summary file. Those overlaps are routing evidence and still need direct page-image comparison before child pages are promoted.
  • Blocked — FBI Vault comparison: The WAR.GOV manifest's fewer-redactions/missing-pages statement is not resolved here. A page-by-page comparison against the FBI Vault copy is still required before Open Sky can publish exact declassification deltas.
  • Partial — debris and specimen photo custody: Rendered pages 22 and 167 confirm that this PDF contains exhibit-style photographs of debris/fragments and alleged specimens. The current packet does not by itself supply the original photo logs, enclosure indexes, or lab records needed to identify every collector, test, or custody handoff for pages 22-25 and 167.
  • Partial — Maury Island/Tacoma timeline: The source text separates Dahl/Crisman statements, Arnold/Smith activity, Palmer/Venture Press correspondence, the B-25 crash record, press-call material, and FBI/Army narrative summaries. The existing Maury Island Incident page is useful context, but this Section 3 packet still needs a source-level timeline before its strands are merged or compared.
  • Partial — Muroc prosaic checks: Section 3 preserves the Muroc witness statements and prosaic leads such as aircraft activity, weather-balloon comparisons, altitude estimates, and test context. Current linked corpus checks found companion Muroc hits in the incident-summary file and 62-HQ-83894_SUB_A; aircraft, balloon, weather, sun-angle, and test-schedule checks remain external/source tasks.
  • Checked — Twin Falls visual distinction: Page 128 is a newspaper artist's conception, while page 133 is an A. C. Urie handwritten sketch/statement page. Both are report depictions preserved in the file, not direct object photographs; any child page should keep the teletype, newspaper art, and Urie sketch as separate evidence types.
  • Partial — Phoenix/Rhoads photograph lead: Section 3 mentions William Rhoads/Rhodes and a request for interview/photos, and the current linked corpus has related hits in Sections 1 and 2. The actual photographs, negatives, and custody trail are not established by this page and still need source retrieval.
  • Checked — Danforth/Wright Field specimen context: Page 167 confirms the alleged-specimen exhibit photograph, while pages 164-165 state the pictured/related specimens had no connection with MOGUL or an AMC project. Keep that as a source-text constraint and custody lead, not as a broad resolution of every Danforth-related reference.
  • Checked — Minot graph crosslinks: The current candidate crosslinks around "Minot" are anchored to Kenneth Arnold biographical text about schooling in Minot, North Dakota. They are not incident matches for this Section 3 packet and should remain audit-only.

Deep investigation — graph + web reconnaissance

Source reread and visual evidence

A second pass through the OCR text and selected page renders reinforces that this page should stay modeled as a multi-thread FBI/Army Air Forces packet rather than a single case. The early F. M. Johnson / Mount Adams material preserves a report near the Kenneth Arnold date, but the Fourth Air Force transmittal itself cautions that Johnson may have read Arnold coverage in newspapers before writing his account. The Guam extract records three enlisted men reporting small crescent-shaped objects at Harmon Field on August 14, 1947, while the West Rindge memorandum records a prosaic lab result: ordinary cast iron subjected to high heat.

The Muroc pages remain the packet's strongest cluster for structured follow-up because they contain named military witnesses, dates, base context, aircraft activity, altitude/speed estimates, and explicit weather-balloon comparisons. Those records include Colonel Gilkey's statement that he believed his object was vapor and did not consider it important enough for further reporting; Major Shoop's thin metallic object seen for about eight minutes; T/Sgt. Ruvolo and Lt. McHenry's silver disc/spherical objects; Major Wise's yellowish-white sphere while running up an XP-84; and Captain Stapp's observation during an ejection-seat experiment with P-82/A-26 aircraft overhead. These are source-text testimony records and prosaic-check starting points, not findings.

Rendered-page review confirms the visual evidence types: pages 22-25 are confidential exhibit-style photographs of dark debris/fragments; page 128 is a Twin Falls newspaper artist's conception with side, bottom, and end-view drawings; page 133 is an A. C. Urie handwritten sketch/statement; and page 167 is a photograph labeled "FLYING SAUCER", ALLEGED SPECIMENS. None of those renders is direct aerial-object imagery. The Danforth/Wright Field text around pages 162-165 is an important control: the visible memo says the alleged specimens had no connection with MOGUL or any Air Materiel Command project and were considered hoax-associated by that command.

Graph connections and audit cautions

Read-only graph checks found two official document records for this asset family: the WAR.GOV release-row record and the PDF asset record. The PDF asset record preserves the official URL, file size, full-download SHA-256, Release 01 provenance, and OCR completion metadata. It is connected to the official release source, the release-row document, a derived UFO-USA markdown conversion, OCR/manifest chunks, 1,159 machine-extracted claim records, 512 entity-mention records keyed to this asset, and 42 machine-extracted sensor-event records.

Those semantic records are scaffolding only. Sample claim and sensor-event records correctly surface high-signal source phrases from Maury Island/Tacoma, Muroc, aircraft-test, balloon, and control-tower passages, but they are marked machine_extracted_needs_human_review and do not carry page-level judgment by themselves. The top extracted entity mentions are dominated by broad dates, FBI, aircraft, reflection, and balloon, which is consistent with a mixed historical packet and also shows why child-cluster splitting is necessary before cross-case analytics.

The graph's 10 CANDIDATE_CROSSLINK edges should remain audit-only. The strongest example is the Minot set: NARA Project Blue Book Minot records, FBI Vault pages, and Larry Hatch database entries are being connected because Arnold's biographical text mentions schooling in Minot, not because Section 3 contains a Minot incident. That is useful for false-positive cleanup and index hygiene, but it is not a provenance, identity, or causality connection.

Official/archive reconnaissance

Live official-source checks were limited by access behavior but useful for custody triage. The WAR.GOV landing page, CSV, press release, and source PDF remain the source-of-record URLs for this asset, though live requests from this verification environment returned access-denied responses; the Open Sky release-file copy therefore remains the verified local public route for byte/hash checks. Direct FBI Vault UFO Part 02 and UFO Part 04 PDF download endpoints responded as official PDFs, and archived snapshots also exist; those are the right comparison targets for the manifest statement that the FBI Vault version is partial or more redacted.

NARA Catalog pages for the graph-linked Minot candidates were reachable, but the graph context shows those are keyword/biographical false-positive leads for this packet. They should not be cited as Section 3 corroboration unless a separate child page establishes a real shared incident anchor. For the Muroc, Twin Falls, Danforth/Wright Field, Seattle/Boeing, and Maury Island/Tacoma clusters, the next official/archive lane is not broad web search; it is targeted retrieval of companion official records, FBI Vault page ranges, NARA/Project Blue Book file units, Air Force historical materials, and any lab/photo enclosure logs that can identify the specimen/debris photographs.

Prosaic checks and follow-up leads

The file itself already contains several prosaic leads that should be modeled before escalation: newspaper-contamination risk for Johnson; ordinary cast iron for West Rindge fragments; aircraft activity, aircraft-test schedules, parachute/ejection-seat context, balloon comparisons, sun reflection, and wind direction for Muroc; weather-balloon releases for the Nichols high-altitude Seattle report; burnt paper blown across Boeing Field for at least one Armstrong-related object; and the Danforth/Wright Field statement that the alleged specimens were not MOGUL or AMC project material.

Priority follow-up is to split Section 3 into child pages, beginning with Muroc, Twin Falls/Urie, Maury Island/Tacoma, Danforth/Wright Field specimens, and Seattle/Boeing. Each child should keep testimony, official agency statements, media/exhibit images, machine-extracted graph claims, prosaic leads, and unresolved questions separated. This page should continue to avoid a Finding/Hypothesis/ResolutionDecision posture until those child clusters have source-level timelines and external official checks.

Audit note

This deep-investigation pass used the current wiki page, OCR text, rendered PDF page images, read-only graph queries, and official/archive URL probes. It added no graph writes and makes no resolution claim.

Limits

This page is a document-packet reading, not a case resolution. No finding, hypothesis, or explanation is asserted here.

The PDF is a mixed archival bundle. It includes original-looking letters, copies, teletypes, official memoranda, report extracts, newspaper clippings, photographs of physical material, reverse sides, and routing sheets. That mixture makes it easy to blend separate events if the packet is summarized too aggressively.

OCR is good enough for page-level navigation but not perfect. Some pages are reversed, faint, damaged, handwritten, or mostly photographic. Rendered page checks confirmed important visual distinctions, but they do not replace a full manual archival transcription.

Visible photographs in this packet are primarily debris/specimen/exhibit photographs. The Twin Falls pages include sketches and newspaper art. The page 188 memorandum references attached photographs of reports, but the rendered page itself does not show those photographs.

The graph context is machine-extracted and remains review material. Extracted claims, sensor events, entity mentions, and candidate crosslinks should be audited against the page images and OCR before being used in public analysis beyond this draft.

Sources