Compiled and edited by Bob Brandenburg, TIGHAR #2286
and
Ric Gillespie, Executive Director

In the days following Earhart’s disappearance on July 2, 1937, government and professional civilian radio operators in the Central Pacific region and on the United States Pacific coast, as well as radio amateurs and shortwave listeners in the continental United States, Hawaii, Canada and Australia reported receiving radio signals that could have been sent from the missing aircraft. Lockheed representatives authoritatively stated that transmissions from Earhart’s Electra (NR16020) were possible only if the airplane was on land. Much of the Coast Guard/Navy search was driven by signals that were considered to be genuine. However, when the search failed to find any trace of the airplane or crew, the radio signals were dismissed by authorities as having been hoaxes or misunderstandings.

Clearly, if any reported signal was genuine, the Electra did not go down at sea. A comprehensive and detailed catalog and analysis of the reported post-loss radio signals must, therefore, be an element in any informed investigation of the Earhart disappearance.

A preponderance of archival and physical evidence unrelated to the post-loss radio signals supports a hypothesis that the Earhart flight ended with a landing on the reef at Gardner Island (now Nikumaroro, part of the Republic of Kiribati), an uninhabited coral atoll in the Phoenix Group. The catalog and analysis presented here tests that hypothesis by evaluating the credibility of each reported signal as having being sent from NR16020 at Gardner.

Radio Frequency Allocation and Usage

In June 1930, the Federal Radio Commission (FRC)^[1] assigned 278 kHz as the frequency to be used for all ground station transmissions to itinerant aircraft (all aircraft other than those flying regular schedules on fixed routes); assigned 3106 kHz as the national calling and working frequency for itinerant aircraft; and authorized 3106 kHz for optional use by transport aircraft in addition to their assigned route frequencies. An aircraft would conduct a two-way voice (A3 mode) conversation with a ground station by transmitting on 3106 kHz, and receiving on 278 kHz.

In October 1930, the FRC^[2] redesignated 278 kHz as the “airport frequency” and expanded its mandated use for ground station transmissions to include all aircraft, not just itinerants.

In December 1931, the FRC^[3] revised the national aircraft calling and working frequency from 3106 kHz to 3105 kHz.

In May 1937, the Federal Communications Commission, successor to the FRC, issued revised regulations^[4] allowing all aircraft to use 6210 kHz as a day-only alternate for 3105 kHz. The revised regulations also: defined a ground station transmitting on 278 kHz as an “airport station;” required such stations to maintain a continuous listening watch on 3105 kHz; and restricted the allowed power on 278 kHz, thus limiting communications with aircraft to short distances – about 30 miles.

Other nations using 6210 kHz^[5] (A3) were: Canada, for aircraft use; the Soviet Union for interior communications; and Venezuela, for broadcasting. 3105 kHz also was used by Canadian aircraft flying the route^[6] between Vancouver and Seattle.

All U.S. pilots were required^[7] to have a third class radiotelephone operator’s license, which authorized operating – but not adjusting – a voice-modulated (A3) transmitter.

The United States was the only nation using 3105 kHz as an aircraft calling and working frequency. In 1937, there were 44 U.S. airport stations^[8] transmitting on 278 kHz and receiving on 3105 kHz; 21 of those stations were west of the Mississippi rive, 14 of which were in Washington, Oregon, and California. There were no such stations in the central Pacific.

In summary, at the time of the Earhart disappearance, by international agreement, the only legal voice radio transmissions on 3105 kHz anywhere in the world were sent by U.S. registered civil aircraft calling a limited number of airports in the continental United States, and Canadian commercial carriers flying between Vancouver and Seattle. The sole exceptions were Amelia Earhart, who had permission to use 3105 kHz as a calling frequency during her world flight, and the Coast Guard cutter Itasca whose calls were duly recorded in the ship’s radio log.

Post-Loss Radio Signal Frequencies

Earhart used 3105 kHz as her “nighttime” frequency and 6210 kHz as her “daytime” frequency. With two exceptions, all reported post-loss radio signals were heard during nighttime at Gardner. Those nighttime signals were heard on 3105 kHz or on the second, fifth or sixth harmonic of that frequency. On the two occasions when credible signals were heard during Gardner daytime, both were heard on 24840 kHz, the 4th harmonic of 6210 kHz, Earhart’s daytime frequency. That frequency is also the 8th harmonic of 3105 but transmitter output at such a high harmonic is virtually nil. There is, therefore, a clear correlation between the frequency upon which credible post-loss signals were heard, the time of day on Gardner Island, and Earhart’s known use of her two primary frequencies.

In some cases, apparently identical signals were heard simultaneously at widely separated locations.

Possible Post-Loss Signal Sources

Earhart transmitted A3 (voice) on 3105 kHz at night, and 6210 kHz during daylight, using her 50-watt WE-13C transmitter. The Coast Guard cutter Itasca, stationed at Howland Island to support the Earhart flight, transmitted A3 on 3105 kHz, but did not have voice capability on 278 kHz or 6210 kHz.

Under favorable propagation conditions, it was possible for aircraft operating on the U.S. west coast at night to be heard on 3105 kHz in the central Pacific. The Itasca reported hearing such signals on one occasion.

There were three 50-watt Morse code radio stations in Nicaragua, on 3102 kHz, 3105 kHz, and 3107 kHz respectively, which could be heard on a receiver tuned to 3105 kHz with bandwidth set to receive voice signals but the stations sent only code, not voice.

No central Pacific ground stations^[9] transmitted on 278 kHz or received on 3105 kHz. All transport aircraft in the area used assigned route frequencies, instead of 3105 kHz. Therefore, other than Itasca, Earhart’s Electra was the only plausible central Pacific source of voice signals on 3105 kHz.

Earhart’s Western Electric model 13C transmitter^[10] had no harmonic suppression circuitry, and thus generated and delivered power to the plane’s dorsal antenna at harmonics of the fundamental frequency – 3105 kHz or 6210 kHz. The antenna’s radiation efficiency^[11] at harmonic frequencies enabled signal reception at distances of thousands of miles under suitable propagation conditions. Therefore, reports of Earhart signals heard on harmonic frequencies cannot be dismissed on the basis of frequency alone.

Were there hoaxes?

The possibility of a hoax must be considered when evaluating a post-loss signal report. A hoax could be perpetrated in two ways: either by transmitting a false signal; or by making a false report.

A false-signal hoax could succeed only if it was heard by someone, and therefore would probably be sent on a frequency which the hoaxer might reasonably expect someone to be monitoring, i.e. 3105 kHz. In some cases, credible sources reported hearing dashes on 3105 kHz, immediately following requests by radio station KGMB in Honolulu for Earhart to send dashes. Those dashes could not have been hoax responses originating in the continental U.S., because KGMB could not be heard^[12] there due to interference from broadcast stations operating on the same frequency as KGMB, in Idaho and Colorado. In other cases, credible sources in widely separated locations in the U.S., Canada, and the central Pacific, reported hearing a woman, speaking English – and in some instances saying she was Amelia Earhart – requesting help on 3105 kHz or a harmonic of that frequency. In one case, two widely separated credible sources, one in Wyoming and one in eastern Canada, simultaneously heard a woman saying she was Amelia Earhart and requesting help. The Wyoming reception was on 15525 kHz, the 5^th harmonic of 3105 kHz, and the eastern Canada reception was on 18630 kHz, the 6^th harmonic of 3105 kHz.

The only known aircraft transmitters capable of transmitting on 3105 kHz were on U.S. and Canadian aircraft. Acquiring such a transmitter – or modifying existing equipment to transmit on 3105 kHz – in time to transmit hoax signals during the few days when post-loss signals were heard would be a daunting task for even the most technically talented would-be hoaxer.

The existence of the frequency band 3500 kHz to 4000 kHz, allocated for amateur (ham) radio operators, might suggest the possibility of ham hoax transmissions on 3105 kHz. However, in addition to the challenge of modifying equipment to operate on 3105 kHz on short notice, a ham would risk federal penalties for transmitting outside the assigned band limits.

But even if a hoaxer had a suitable transmitter, and was a woman or had a female accomplice, it would be impossible to control who would hear the signal, and thus impossible to direct the hoax to a specific target or group of targets. A hoax transmitter in the continental U.S. (CONUS) should have been heard by at least one of the 44 airport stations maintaining a continuous listening watch on 3105 kHz; a hoax transmitter on or near the west coast should have been heard by the special Coast Guard facility set up near San Francisco to listen for Earhart signals; and a hoax transmitter in Hawaii should have been heard loud and clear simultaneously at the Navy, Coast Guard, and Pan American Airways stations in Hawaii listening for Earhart signals. No such signals were reported.

A hoaxer elsewhere in the central Pacific not only would need a suitable transmitter, but also would need to adjust the transmitter’s modulation – voice or dashes – and output power to produce the distinctive signal characteristics reported by listeners in the central Pacific, Canada, and CONUS. And finally, a hoaxer responsible for the signals which directional bearings indicate originated in the Phoenix Islands would need time to position himself in that remote corner of the Pacific – and therefore must have had advance knowledge that Earhart was not going to reach Howland Island.

Given the numerous constraints militating against successfully perpetrating a signal transmission hoax, the likelihood of such events is vanishingly small.

However, known hoax reports did occur. Several of the post-loss signal reports analyzed in this catalog were determined to be hoaxes. These were not reports from people who heard hoax transmissions but rather reports from people who, for whatever reason, claimed to have heard something they did not hear.

Post-loss Radio Signal

A radio signal reported to have been received after 2330Z , July 2, 1937 (Local noon aboard the Coast Guard cutter Itasca) – when the Coast Guard assumed that NR16020 was down – and with characteristics or content suggesting the signal was, or could have been, sent from the aircraft.

Reception Probability

The probability that a reported signal, if sent from NR16020 at Gardner, would be heard at the receiver site. This probability is computed from the signal statistics reported by the Ionospheric Communications Enhanced Profile Analysis and Circuit Prediction Program (ICEPAC).^[13]

Qualitative Factors

Aspects of the reported signal affecting its credibility. For example, the presence of occult information – content or characteristics consistent with a genuine signal from NR16020 but unknowable as such by the receiving party – is a strong positive factor; the degree to which reported signal characteristics – fading, understandability, etc. – agree with ionospheric propagation conditions can be a positive or negative factor.

Morse code proficiency in the transmission of a message is a strong negative factor. The Itasca logged numerous instances of hearing Morse signals on 3105 kHz, but neither Earhart nor her navigator, Fred Noonan, was proficient in Morse code. They presumably could compose a Morse message, character-by-character, using a Morse code table to get the dot/dash symbol string for each English character, and then tap out the message by using the transmitter microphone push-to-talk button. But that process would be slow and error-prone, and would not come close to approximating the pace and rhythm of a trained Morse operator. Itasca also sent Morse messages asking Earhart to perform various actions such as to send dashes in response or state her geographic position. Earhart and Noonan could recognize single letters if sent slowly and repeatedly,^[14] but understanding the content of a Morse message, sent at even moderate speed, was out of the question, and any presumed responses to such requests would be implausible.

A long duration carrier signal – in which the transmitter is keyed, i.e. “on the air,” but no voice modulation is present – for several minutes or more, is a strong negative factor. There was no plausible reason for such a transmission.

Credibility

A determination as to whether the evidence supports a conclusion that a signal was sent from NR16020. A signal is rated as Credible if it was heard on 3105 kHz, 6210 kHz, or a harmonic, and positive qualitative factors were present. A signal is rated Not Credible if there are factors precluding a finding of credible. If there is insufficient evidence to decide whether a signal is credible or not credible, it is rated Uncertain.

The known post-loss transmissions to Earhart and reported post-loss signals from Earhart are listed chronologically. Each signal is cataloged in a table comprising twelve parameters: