MYS0CYPMFQR4AWT·Mystery · disappearance

The disappearance of Amelia Earhart

Earhart and Noonan left Lae on the morning of 2 July 1937 for Howland Island, a strip of coral two miles long where the United States had built a runway and stationed the Coast Guard cutter Itasca to guide them in. The Itasca heard Earhart's voice with increasing strength through the morning but could not establish two-way contact: she could not hear the cutter's voice transmissions, and her direction-finding loop could not take a bearing on its signal. At 07:42 she reported "we must be on you but cannot see you, gas is running low"; at 08:43 she gave her last confirmed message, that she was "on the line 157 337" and "running north and south," the sun line Noonan would have advanced to pass through Howland⁠[1]⁠[2]. Nothing certain was heard again.

The Navy and Coast Guard searched for sixteen days with the battleship Colorado, the carrier Lexington and their aircraft, covering some 250,000 square miles of ocean and the islands within the Electra's reach. On 9 July Colorado's floatplanes flew over the Phoenix Islands; at Gardner Island the pilot saw "signs of recent habitation" but no aircraft and no one answering his circling⁠[3]. The search ended on 18 July, and the official position was that the aircraft had come down at sea near Howland⁠[2].

Two things have kept the question open. The first is a set of radio signals heard on Earhart's frequencies in the days after the loss, by Pan American Airways direction-finding stations, Navy and Coast Guard operators and amateurs in the Pacific and North America; some were hoaxes and misidentifications, and a few are hard to dismiss⁠[1]. The second is Gardner Island itself, uninhabited in 1937 and settled in 1938, where in 1940 a colonial officer found a partial skeleton, a sextant box, and the remains of a woman's shoe beneath a tree⁠[4]. A Fiji physician measured the bones in 1941 and judged them a stocky male; the bones were then lost⁠[5]. A 1998 re-analysis of his measurements said female and European; a 2015 review said he had been right; a 2018 study by the forensic anthropologist Richard Jantz found that the bones were more similar to Earhart than to 99 per cent of a reference sample⁠[6]⁠[7]⁠[8]. The International Group for Historic Aircraft Recovery (TIGHAR) has mounted a dozen expeditions to the island since 1989, Robert Ballard searched its reef slope in 2019, and in October 2026 a Purdue-backed team went to test a satellite anomaly in the lagoon; all returned without the aircraft⁠[2]⁠[9]. Of the 3 questions on this site, 0 are answered. The most likely answer to “Where did the Electra come down?”, not yet confirmed: it ran out of fuel and ditched at sea within about a hundred miles of Howland, and sank.

Key questionsset by the editors

Where did the Electra come down?Most likely: it ran out of fuel and ditched at sea within about a hundred miles of Howland, and sank

Whose bones were found on Gardner Island in 1940?No answer ahead yet

All 3 questions

What is known

The Itasca's radio log is the primary record of the morning. It shows Earhart transmitting on schedule from 02:45 ship time, her signals rising to strength five by 07:42, the cutter transmitting voice and Morse that she did not acknowledge, and the failure of both direction-finding attempts, hers because the Itasca's transmissions were too short and on the wrong frequency for her loop, theirs because the experimental high-frequency finder on Howland could not get a bearing on her brief voice calls⁠[1]⁠[2]. The last message at 08:43 was heard at strength five, which implies the aircraft was within perhaps a hundred miles of the cutter⁠[1]. Fuel estimates for the Electra at that hour run from almost none to several hours, depending on the power settings assumed for the night; the published analyses disagree, and the question is unsettled⁠[2].

Lockheed 10-E Electra NR16020Smithsonian Institution (photographer not recorded), 1937. Public domainArtifact

The post-loss radio record is large and mixed. TIGHAR's catalogue lists 182 reported receptions between the evening of 2 July and 10 July on 3105 and 6210 kHz or their harmonics, rated for credibility on propagation and content; the Pan Am stations at Wake, Midway and Mokapu Point took bearings on an unidentified carrier on several nights that crossed in the vicinity of the Phoenix Islands⁠[1]. The Navy acted on them: the Colorado's search of the Phoenix group followed the bearings. None of the signals is proven to be Earhart's, and the catalogue is compiled by the organisation that holds the Nikumaroro hypothesis, which readers should weigh⁠[1].

The Gardner Island record begins with the 1937 overflight and resumes with the 1940 finds. Gerald Gallagher, the Phoenix Islands settlement officer, reported to his superiors in Fiji a skull found by labourers, then a partial skeleton beneath a ren tree on the south-east part of the island, with a sextant box, part of a woman's shoe and a Benedictine bottle, and wrote that there was "a very slight chance" they were Earhart's⁠[4]. D. W. Hoodless at the Central Medical School, Suva, examined thirteen bones on 4 April 1941, took measurements, and concluded they were a male of European or mixed ancestry, about 5 feet 5 inches tall⁠[5]. The bones and the box have not been found since. Everything later said about the bones rests on his measurements.

Map of the worldOpen the full map

+−

Howland IslandIsland

Mapped

Nikumaroro (Gardner Island), Phoenix Islands, KiribatiAtoll

Within 1 km

Jaluit Atoll, Marshall IslandsAtoll

Within 10 km

Central Medical School, Suva, FijiInstitution

Within 10 km

MappedWithin 1 kmA regionNatural Earth

Timeline

Timeline

Squeezed where little happened. Each mark is a published claim, by how it stands:AcceptedMost likelyStill possibleDisputedRuled out

The search

The crash-and-sink reading has been tested by sonar. Nauticos, with the navigator Elgen Long's fuel and navigation analysis, searched the deep seabed north-west of Howland in 2002, 2006 and 2017; the Waitt Institute searched in 2009; together they covered about 3,610 square miles at depths near 5,000 metres without finding the aircraft⁠[10]. In January 2024 Deep Sea Vision published a sonar image of an aircraft-shaped object about 100 miles from Howland; a return visit in November 2024 showed it to be a rock formation⁠[11]⁠[12]. In August 2025 Nauticos announced a new analysis of the 1937 radio receptions using a restored Electra radio set, which it said narrowed the search area, and a fourth expedition for which it was raising funds; as of October 2026 it had not sailed⁠[10].

The Nikumaroro reading has been tested on land and on the reef. TIGHAR's expeditions since 1989 recovered artefacts at the "Seven Site," where the bones were found: a jar of the kind that held freckle cream, a compact, a zip fastener, glass, fish and bird bones from campfires, and elsewhere on the island a sheet of aircraft aluminium; none carries a serial number that ties it to the Electra, and the island was inhabited from 1938 to 1963⁠[2]. A 1937 photograph of the reef taken by the British officer Eric Bevington shows an object that TIGHAR read as landing gear; a 2019 survey found rocks of that shape on the reef⁠[2]. Ballard's 2019 expedition with National Geographic mapped the reef slope and found no wreckage⁠[13]. In 2020 a satellite image of the lagoon beside the Taraia peninsula showed a reflective object of the Electra's size; the Archaeological Legacy Institute and the Purdue Research Foundation, Earhart's former employer and the Electra's original owner, planned an expedition for November 2025, postponed it to 2026, and went in October 2026. The object was a cluster of large tree trunks with a shiny coating of carbonate-cemented coral sand; one Y-shaped segment remained. The team also recovered a steel tube about 30 inches long from the lagoon and said the search would continue⁠[9]⁠[14].

The explanations

Crash and sink. The aircraft ran out of fuel near Howland, within the area the last signal strength implies, and sank in deep water. This is the simplest reading of the radio log and was the Navy's conclusion⁠[2]. It predicts that no credible post-loss signals exist, because the Electra's transmitter could not operate from the water, and that the wreck lies on the seabed within about a hundred miles of Howland. Against it are the post-loss receptions that the catalogue rates credible and the Pan Am bearings; for it are the 3,610 square miles searched without result, which however cover only a part of the possible area⁠[1]⁠[10].

Nikumaroro. Unable to find Howland, Earhart flew south-east along the 157–337 line, reached Gardner Island, landed on the reef flat at low tide, transmitted for several nights until the aircraft was taken by the surf, and died on the island as a castaway⁠[2]. The reading predicts credible post-loss signals timed to low tides, a castaway's remains and belongings on the island, and aircraft wreckage on the reef slope. The first is argued from the catalogue; the second from the 1940 finds and the Seven Site artefacts; the third has not been found by Ballard's survey or by TIGHAR's own⁠[1]⁠[4]⁠[13]. The bones are the crux. Hoodless said male; Burns and colleagues, from his measurements, said a tall European woman; Cross and Wright, reviewing both, said Hoodless's first-hand examination was the more robust and the bones were a robust middle-aged man; Jantz, using Earhart's own measurements from clothing and photographs against modern reference data, found the bones closer to her than to 99 per cent of the comparison sample⁠[6]⁠[7]⁠[8]. All of them are working from thirteen measurements of bones nobody has seen since 1941. The 1937 overflight that saw "signs of recent habitation" and no one waving is consistent with the reading only if the castaways were dead or unable to signal within a week⁠[3].

Japanese capture. The aircraft came down in the Marshall Islands, and Earhart and Noonan were taken to Saipan and died there. The reading dates from wartime rumour and Fred Goerner's 1966 book, rests on Marshallese and Saipanese recollections gathered decades later, and has produced no document, no wreckage and no grave⁠[2]. The 2017 photograph said to show the pair on a Jaluit dock was found by a Japanese researcher in a travel book published in 1935⁠[15]. The reading is contradicted by the radio log, which has the aircraft near Howland at 08:43, some 800 miles from the Marshalls with little fuel⁠[1].

Published claims

Published claims and their status

Year	Who	Claimed	Status	Reason
1937	John O. Lambrecht, USS Colorado⁠[3]	Gardner Island showed signs of recent habitation, but no aircraft was seen and no one answered the circling floatplanes.	Accepted	Nobody disputes what the pilots reported; what the signs of habitation were is still argued⁠[2]
1940	Gerald Gallagher⁠[4]	There is a very slight chance that the skeleton, sextant box and shoe found on Gardner Island were Earhart's.	Disputed	The bones were lost after 1941 and the analyses of their measurements disagree⁠[7]⁠[8]
1941	D. W. Hoodless⁠[5]	The Gardner Island bones are a male's, of European or mixed ancestry, about 5 feet 5 inches tall.	Disputed	Re-analyses of his measurements since 1998 disagree about the sex and stature⁠[6]⁠[7]⁠[8]
1966	Fred Goerner⁠[16]	Earhart and Noonan came down in the Marshall Islands and were taken by the Japanese to Saipan, where they died.	Disputed	No document, wreckage or grave has turned up in sixty years, and the radio log puts the aircraft near Howland with little fuel at 08:43⁠[2]⁠[1]
1998	Karen Burns, Richard Jantz, Thomas King and Richard Gillespie⁠[6]	Hoodless's measurements fit a tall white woman of northern European ancestry better than a man.	Disputed	A peer-reviewed review found Hoodless's first-hand male identification the more robust⁠[7]
1999	Elgen and Marie Long⁠[17]	The Electra ran out of fuel near Howland Island and sank.	Most likely	It is the simplest reading of the radio log, but about 3,610 square miles of seabed near Howland have been searched by sonar without finding the aircraft⁠[1]⁠[10]
2006	Ric Gillespie (TIGHAR)⁠[18]	Earhart landed on the reef at Gardner Island, transmitted for several nights, and died there as a castaway.	Disputed	Decades of expeditions have found suggestive artefacts but no wreckage, and the reef-slope surveys came back empty⁠[13]⁠[9]
2015	Pamela Cross and Richard Wright⁠[7]	Hoodless's first-hand examination is more reliable than the later re-analyses, and the bones were a robust middle-aged man's.	Disputed	Jantz's 2018 comparison with Earhart's own dimensions reached the opposite result from the same measurements⁠[8]
2017	History (Amelia Earhart, The Lost Evidence)⁠[19]	A photograph in the National Archives shows Earhart and Noonan alive on a dock at Jaluit after their capture.	Ruled out	The photograph was published in a Japanese travel book in 1935, two years before the flight⁠[15]
2018	Richard Jantz⁠[8]	The Nikumaroro bones are more similar to Earhart than to 99 per cent of a modern reference sample.	Disputed	The result rests on thirteen 1941 measurements and on Earhart's dimensions reconstructed from clothing and photographs; Cross and Wright read the same bones as male⁠[7]
2021	Robert Brandenburg and Richard Gillespie (TIGHAR)⁠[1]	Some radio signals heard after the loss were credible transmissions from the Electra, and Pan Am bearings crossed near the Phoenix Islands.	Disputed	The credibility ratings are TIGHAR's own and have not been independently replicated, and the Navy searched the Phoenix Islands on those bearings without result⁠[3]
2024	Deep Sea Vision⁠[11]	A sonar image about 100 miles from Howland shows an object shaped like the Electra on the seabed.	Ruled out	A return visit with cameras in November 2024 showed the object to be a rock formation⁠[12]
2025	Nauticos⁠[10]	A new analysis of the 1937 radio receptions, using a restored Electra radio set, narrows the area near Howland where the aircraft should lie.	Still possible	The expedition to search the narrowed area had not sailed by October 2026⁠[10]
2025	Archaeological Legacy Institute and Purdue Research Foundation⁠[20]	The Taraia Object, a reflective shape in the Nikumaroro lagoon seen in satellite images, may be the Electra.	Ruled out	The October 2026 expedition found it to be a cluster of coated tree trunks⁠[9]⁠[14]
2026	Archaeological Legacy Institute and Purdue Research Foundation⁠[9]	The Taraia Object is a cluster of large tree trunks coated with carbonate-cemented coral sand, not an aircraft.	Accepted	Seen on site by the expedition and reported independently⁠[14]

What would settle it

Wreckage with a serial number. The Electra's engines and structure carried numbers that would identify them on the seabed off Howland or on the reef slope at Nikumaroro, and both places have been searched only in part. Nauticos's refined radio analysis, if it holds up, narrows the Howland field; the lagoon and reef at Nikumaroro have been mapped but the deep slope only along transects⁠[10]⁠[9]. The bones would settle the island question by DNA if they were ever found in Fiji; the search of the medical school's stores has been made more than once without result⁠[2]. A full, independent re-analysis of the post-loss radio receptions, using modern propagation models and the original logs rather than TIGHAR's catalogue, would tell whether any of them could have come from a transmitter on Gardner's reef.