Ghost Rockets

What the file holds

The only account from the time in the US releases is an article of four pages, “‘Ghost Rockets’ over Scandinavia”, in Intelligence Review No. 47 of 9 January 1947, which the Intelligence Division of the War Department General Staff published (DOW-UAP-D099). The file holds the cover, the contents, the article (pp. 17–20 of the issue) and the back cover; the rest of the issue is not in it. No author is named, the article names no witness, and it gives no place more exact than southern or central Sweden.

No Swedish document is in the releases: not a report, not a communiqué of the Defense Staff, not a page of its investigation. What Sweden said is known only as US documents relay it. The rest of what the releases hold about the reports of 1946 comes from later US papers, the latest of them a G-2 paper of 1950, and from a French association’s report of 1999. We searched the text of all six releases for ghost rockets, Sweden, Norway, Finland, Denmark and Scandinavia, allowing for scanning errors; the documents cited here are all that bear on 1946. Reports from Finland, Denmark, Norway and the Swedish coast in early 1948 are in the US Air Force’s incident summaries (DOW-UAP-D087, DOW-UAP-D088); they concern 1948 and are not part of this case.

The reports

The article opens: “Flying missiles were first reported over southern Sweden in late May 1946 by the press, which gave the missiles the name of ‘Ghost Rockets.’ In June, the missiles also had been reported over Finland and Denmark. By July, the number of sightings over Sweden had greatly increased, and several also had been reported over Norway.” (p. 17) The Defense Staff asked the public to report unusual observations, “and by the end of July almost 1,000 such reports had been received” (pp. 18–19).

The review is plain about who made them: “The great majority of these reports were made by untrained observers and, as would be expected, varied widely in the description of the actual missiles as well as, of their course, altitude, and speed.” It goes on:

“The two most common descriptions of the missiles were ‘a ball of fire with a tail’ and a ‘shiny cigar-shaped object.’ The reported direction of flight covered all points of the compass, with a northerly direction being slightly predominant. Variations in altitude ranged from treetop height to 160,000 feet, the higher altitudes almost exclusively being reported from Finland. Speeds reported were from 65 m. p. h. to ‘lightning fast,’ with the majority described as having great or very great speed.” (p. 17)

A few were reported “as diving into the ground or into lakes, or exploding in the air”. Of what was found, the review has one sentence: “In no case have fragments been found other than bits of material described as ‘nonmetallic slag.’” (p. 17) It does not say where the slag was found, by whom, or whether anyone analysed it, and no other document does. Of the sounds, it records that “Three of the people reporting sightings have mentioned the noise of an outboard motor”, and that one of them “stated that what he saw looked like the V-1 he had seen over London during the war” (p. 20).

A pilot’s report of 14 August

The review calls it “One of the few reports available from an individual who should be more reliable and competent than the average layman” (pp. 17–18): the report of a Swedish Air Force pilot, who is not named.

“On 14 August, at 1000 hours, he was flying at 650 feet over central Sweden when he saw a dark, cigar-shaped object about 50 feet long and 3 feet in diameter flying 200 feet above and approximately 6,500 feet away from him at an estimated speed of 400 m. p. h. The missile had no visible wings, rudder, or other projecting part; and there was no indication of any flame or light as has been reported in the majority of other sightings.” (p. 18)

The review did not trust it: “His report states that the missile was maintaining a constant altitude over the ground and, consequently, was following the large features of the terrain. The last statement casts doubt on the reliability of the entire report because of the inability of a missile, without wings, to maintain a constant altitude over hilly terrain.” It adds that the pilot “does imply that there may have been wings which he was unable to see, because he stated that it could not have been a Swedish jet plane as there was none flying in the vicinity at that time” (p. 18).

The press, the censor and the Swedish investigation

“The Scandinavian press, with the exception of the Communist papers, initially reported the incidents in some detail and openly attributed them to missiles fired by the U. S. S. R. In August, a partial censorship was imposed on the press, restricting the publication of exact details or localities where missiles were seen.” The Communist press, the review goes on, ridiculed the matter and charged that the missiles came from the United States (p. 18).

The Swedish authorities began their investigation in June. In the review’s words it “has been carried out by the Swedish Defense Staff in a very peculiar manner. In the beginning, many of the key personnel were on summer leave, and they were not recalled to deal with the problem.” Then:

“Spokesmen for the Defense Staff repeatedly have told the United States Military Attaché that they definitely believed there were rockets over Sweden, and that they were launched by the Soviets from Peenemünde on the German Baltic coast. However, they have not been able to produce any evidence to support these statements. To date, no United States military or naval personnel in Sweden have seen any fragments, points of impact, or other direct evidence to prove that guided missiles have appeared over Sweden.” (p. 19)

The Military Attaché’s own reports are not in the release.

Two communiqués

What the Defense Staff told the press did not match what its officers said in conversation:

“The official communiqués issued to the press by the Defense Staff have not reflected the same tone as the statements made by individual officers in conversations. In fact, the communiqués themselves have varied. The last one, issued on 10 October, very strongly indicated that the great majority of the reported incidents were of celestial origin, which is a complete change from the one of 6 August when it was reported that, except in a few cases, they could not be meteorites.” (p. 19)

Neither communiqué is in the release; both are known only from this summary. The review read the change as politics: “The complete change in attitude between the communiqués issued on 6 August and 10 October indicates that the Swedish Staff was trying to minimize the entire matter which had grown to such a size that the Staff was afraid of its having official repercussions on Swedish-Soviet relations.” (pp. 19–20)

What the US analysts concluded

The review’s conclusions start from what it took to be known: “The Soviets are known to be working on various guided missiles. They have the ability to produce, and probably have tested, missiles of the V-1 type.” It adds: “There is some evidence that such work has been carried out at Stolp (in Polish-administered Germany) or on the Baltic islands of Oesel or Dago”, and does not say what that evidence was. Then:

“The best evidence, at present, is that there have been only 2 or 3 real incidents, perhaps as many as 5 or 10, of low-flying missiles of the V-1 type. The high-altitude missiles reported seem definitely to have been meteors or fireworks. The Swedish Defense Staff probably has taken advantage of the situation for political purposes and allowed the newspapers to make a big story out of the missiles, without admitting that the Staff had any evidence to indicate that there actually were any such missiles. This was done at a time when the Swedish public was demanding reductions in defense expenditures.” (p. 20)

The review does not say which reports it counted among the 2 or 3, or the 5 or 10, nor on what grounds. A page earlier it had written: “To arrive at any definite conclusion from the conflicting evidence available on these reported missiles is impossible.” (p. 19)

What later documents add

• 19 August 1947. An FBI memorandum records what Lieutenant Colonel Garrett of the Air Forces Intelligence told Special Agent S. W. Reynolds: “when flying objects were reported seen over Sweden, the ‘high brass’ of the War Department exerted tremendous pressure on the Air Forces Intelligence to conduct research and collect information in an effort to identify these sightings” (FBI file 62-HQ-83894, Section 2, p. 129).
• 23 January 1948. Air Materiel Command, for its investigation of flying discs, Project “SIGN”, asked the Chief of Staff of the Air Force, for the attention of Lt Col George Garrett, for “the entire file ‘Swedish Incidents’”: “It is our understanding that a comprehensive file of these incidents accumulated during the past two years is located at your Headquarters.” (DOW-UAP-D100, p. 152) The file is not in the release.
• 10 December 1948. Air Intelligence Division Study No. 203 set the Scandinavian reports in a speculation about Soviet propaganda: “It will be remembered that strange objects first appeared over the Scandinavian countries in 1946. The objects observed there had unusual range and unusual performance characteristics.” (DOW-UAP-D094, p. 10) The same study judged it improbable that US reports had been influenced by “previous sightings of unidentified phenomena in Europe, particularly over Scandinavia in 1946” (p. 6).
• 31 January 1949. The FBI office in San Antonio: “The first such sightings were reported from Sweden, and it was thought that the objects, the nature of which was unknown, might have originated in Russia.” It does not say who thought so (Section 4, p. 115).
• May 1950. An unsigned G-2 paper in the FBI’s file, marked as received from G-2 on 22 May 1950, retells the events: newspapers had speculated about Soviet missiles from Peenemunde or Dago Island; “The fact that an experimental V-2 fired by the Germans from Peenemunde had landed in Sweden in 1944 lent credence to their speculation”; the Defense Staff “may have deliberately allowed the public to believe this theory for some time for political reasons”. And: “Careful investigation by the US MA and by an independent British team, failed to uncover any evidence to support the theory of Soviet missiles. By 1947 the thing had snowballed to such an extent that the Swedish Government finally issued a report identifying most of the incidents with natural phenomena and denying that any foreign power was involved.” (Section 5, pp. 156–157) The paper cites no source.
• An undated clipping in the same file: “Skeptical scientists recalled the mysterious ‘rockets’ seen over Sweden last year. Eighty per cent of the ‘ghost rockets’ proved to be meteors, and Swedish officials said the others could be discounted as pure imagination.” (Section 1, p. 53) The page names no paper and no date.
• 1999. A report of the French association COMETA, in an English translation filed in the release, dates the reports “From February to December 1946”, says that “close to a thousand were detected” and that “no debris was ever found (officially)”, and that Soviet tests were long suspected but “This hypothesis has since been completely ruled out.” It gives no source for any of it, and its dates and its “no debris” do not match the review of 1947.

What is explained, and what is not

Three official positions are on record, and none of them reaches us first hand:

• The Swedish Defense Staff, in its communiqué of 10 October 1946: “the great majority of the reported incidents were of celestial origin”. Known only from the US review’s summary (DOW-UAP-D099, p. 19).
• The US analysts, in January 1947: “only 2 or 3 real incidents, perhaps as many as 5 or 10, of low-flying missiles of the V-1 type”; the high-altitude objects “meteors or fireworks” (p. 20).
• The Swedish Government, in a report of 1947 “identifying most of the incidents with natural phenomena and denying that any foreign power was involved”. Known only from the G-2 paper of 1950 (FBI file 62-HQ-83894, Section 5, p. 157).

No Swedish document is in the US releases: not a report, not a communiqué, not the Government’s report. Every Swedish position here is known at second hand.

Outside the releases, in a podcast of Riksarkivet (episode of 27 April 2023, transcript published in 2024), the head of the War Archives says that the Defense Staff’s committee met for the last time on 12 December 1946 and never presented a tenable explanation. Beside it stands the G-2 paper of 1950, which says that by 1947 the Swedish Government issued a report “identifying most of the incidents with natural phenomena and denying that any foreign power was involved”.

Both Swedish statements, as they are relayed, put most of the reports down to celestial or natural causes, and the US analysts agreed for the high-altitude ones. What stays open is the rest: the 2 or 3, or 5 or 10, low-flying objects that the US review took for real missiles without naming them or giving its evidence; the “nonmetallic slag”, which no document says was analysed; and the reports themselves, none of which is in the release. That is why the status is insufficient data, not explained.

Where it was

The review places the first reports “over southern Sweden” and the pilot’s sighting “over central Sweden”, and names Finland, Denmark and Norway as the other countries with reports; the highest altitudes came “almost exclusively” from Finland (p. 17). It gives no place more exact. A footnote adds: “In September and October, flying objects were reported over widely separate points in Europe and Africa, including Belgium, Greece, Italy, Morocco, and Austria. In the main, these reports have not been confirmed and appear to be due to explainable causes, such as meteors or Very flares, fireworks, etc.” (p. 19)

The pin is OpenStreetMap’s point for Sweden (relation 52822), rounded by us to 59.67° N, 14.52° E. It marks the country, not the place of any sighting.

Why this scores where it does

Witnesses: 2. Almost 1,000 reports, “the great majority” from untrained observers, none of them named. The one report from a trained observer, the Swedish Air Force pilot of 14 August, is the one the review says “casts doubt on the reliability of the entire report”.

Radar and optics: 0. No document in the releases mentions radar in connection with the reports of 1946, and none mentions a photograph or a film.

Physical trace: 1. “Bits of material described as ‘nonmetallic slag’”, with no analysis, no place and no names; and no US serviceman in Sweden saw “any fragments, points of impact, or other direct evidence”.

Multi-channel: 2. The press, the Defense Staff’s statements and its officers’ conversations with the US Military Attaché all reach us through one US review; the later US papers add retellings, not new observations.

Documentation: 3. One summary from the time, four pages, dated and specific about the period, and later papers that add two facts: a report of the Swedish Government, and the negative result of the US and British inquiries. The Swedish files, the communiqués, the Military Attaché’s reports and the file “Swedish Incidents” are not in the release.

Anomaly: 1. The review puts the high-altitude reports down to meteors or fireworks and the low-flying rest to missiles of the V-1 type; what is left is unidentified, not anomalous in any way a document measures.

Data today: 1. Only summaries can be read today; none of the reports, and nothing Swedish, is in the release.

The overall is 1.2: class D, testimony only.

What would settle it

The releases themselves show that two of these records existed:

• The file “Swedish Incidents”, which Air Materiel Command understood in January 1948 to be held at Air Force headquarters (DOW-UAP-D100, p. 152).
• The findings of the “independent British team” and of the US Military Attaché, which the G-2 paper of 1950 mentions (FBI file 62-HQ-83894, Section 5, p. 157).

And beyond them:

• The Swedish Government’s report of 1947, and the Defense Staff’s communiqués of 6 August and 10 October 1946.
• The Defense Staff’s files of the reports, the pilot’s report of 14 August among them.
• An analysis of the “nonmetallic slag”.
• The evidence behind the review’s “2 or 3 real incidents, perhaps as many as 5 or 10”, and behind its “some evidence” of work at Stolp, Oesel or Dago.