On the morning of July 2, 1937, Amelia Earhart and her navigator, Fred Noonan, took off from Lae, New Guinea, on one of the last legs in their historic attempt to circumnavigate the globe. Their next destination was Howland Island in the central Pacific Ocean, some 2,500 miles away. A U.S. Coast Guard cutter, the Itasca, waited there to guide the world-famous aviator in for a landing on the tiny, uninhabited coral atoll.

But Earhart never arrived on Howland Island. Battling overcast skies, faulty radio transmissions and a rapidly diminishing fuel supply in her twin-engine Lockheed Electra plane, she and Noonan lost contact with the Itasca somewhere over the Pacific. Despite a search-and-rescue mission of unprecedented scale, including ships and planes from the U.S. Navy and Coast Guard scouring some 250,000 square miles of ocean, they were never found.

In its official report at the time, the Navy concluded that Earhart and Noonan had run out of fuel, crashed into the Pacific and drowned. A court order declared Earhart legally dead in January 1939, 18 months after she disappeared. From the beginning, however, debate has raged over what actually happened on July 2, 1937 and afterward. Several alternate theories have surfaced, and many millions of dollars have been spent searching for evidence that would reveal the truth of Earhart’s fate.

In her last radio transmission, made at 8:43 am local time on the morning she disappeared, Earhart reported flying “on the line 157 337...running north and south,” a set of directional coordinates that describe a line running through Howland Island.

In 1989, an organization called the International Group for Historic Aircraft Recovery (TIGHAR) launched its first expedition to Nikumaroro, a remote Pacific atoll that is part of the Republic of Kiribati. TIGHAR and its director, Richard Gillespie, believe that when Earhart and Noonan couldn’t find Howland Island, they continued south along the 157/337 line some 350 nautical miles and made an emergency landing on Nikumaroro (then called Gardner Island). According to this theory, they lived for a period of time as castaways on the tiny, uninhabited island, and eventually died there.

U.S. Navy planes flew over Gardner Island on July 9, 1937, a week after Earhart’s disappearance, and saw no sign of Earhart, Noonan or the plane. But they did report seeing signs of recent habitation, though no one had lived on the atoll since 1892.

In 1940, British officials retrieved a partial human skeleton from a remote part of Nikumaroro; a physician subsequently measured the bones and concluded they came from a man. The bones themselves were later lost, but TIGHAR analyzed their measurements in 1998 and claimed that in fact they most likely belonged to a woman of European ancestry, of around Earhart’s height (5-foot-7 to 5-foot-8). In 2018, a forensic analysis of the bone measurements conducted by anthropologists from the University of Tennessee (in cooperation with TIGHAR) showed that “the bones have more similarity to Earhart than to 99 percent of individuals in a large reference sample,” according to a university statement at the time.

A competing theory argues that when they failed to reach Howland Island, Earhart and Noonan were forced to land in the Japanese-held Marshall Islands. According to this theory, the Japanese captured Earhart and Noonan and took them to the island of Saipan, some 1,450 miles south of Tokyo, where they tortured them as presumed spies for the U.S. government. They later died in custody (possibly by execution).

Since the 1960s, the Japanese capture theory has been fueled by accounts from Marshall Islanders living at the time of an “American lady pilot” held in custody on Saipan in 1937, which they passed on to their friends and descendants. Some of the theory’s advocates suggest that Earhart and Noonan were in fact U.S. spies, and their around-the-world mission was a cover-up for efforts to fly over and observe Japanese fortifications in the Pacific. At the time, more than four years before the Pearl Harbor attack, Japan was not yet the Americans’ enemy in World War II.

Some have suggested that Earhart didn’t die on Saipan after her capture, but was released and repatriated to the United States under an assumed name. Beginning in the 1970s, some proponents of this theory have argued that a New Jersey woman named Irene Bolam was in fact Earhart. Bolam herself vigorously denied these claims, calling them “a poorly documented hoax,” but they persisted even long after her death in 1982.

Since 1989, TIGHAR has made at least a dozen expeditions to Nikumaroro, turning up artifacts ranging from pieces of metal (possibly airplane parts) to a broken jar of freckle cream—but no conclusive proof that Earhart’s plane landed there.

Amid ongoing controversy, spanning more than 80 years of debate among researchers and historians, the crash-and-sink theory remains the most widely accepted explanation of Earhart’s fate. But over three expeditions since 2002, the deep-sea exploration company Nauticos has used sonar to scan the area off Howland Island near where Earhart’s last radio message came from, covering nearly 2,000 square nautical miles without finding a trace of the wreckage of the Electra. Until that wreckage—or some other definitive piece of evidence—is found, the mystery surrounding Amelia Earhart’s final flight will likely endure.