Tracing Earhart’s Unfinished Journey Around the World

Amelia Earhart’s repaired 1937 world-flight attempt traveled eastward near the equator in a Lockheed Model 10E Electra, with Fred Noonan as navigator. Its completed path can be traced from Oakland across the southern United States, the Americas, Africa, Asia, Indonesia, and northern Australia to Lae, New Guinea.

The evidence changes after Lae. Earhart and Noonan departed for Howland Island on July 2, 1937, but never arrived. Howland, Honolulu, and Oakland formed the intended finish—not the completed route. Radio transmissions were received during the final flight, but they do not establish the Electra’s precise track or final location according to the National Air and Space Museum’s account.

The essential distinction for any Amelia Earhart flight path is therefore between what was completed, what was attempted, what remained unflown, what was reported by radio, and what is still unknown.

The 1937 route at a glance: completed, planned, and unknown

The broad completed route was:

Oakland → Miami → northeastern South America → Brazil → South Atlantic → West and central Africa → South and Southeast Asia → Indonesia → northern Australia → Lae, New Guinea

The attempted final stage was:

Lae → Howland Island

The intended but never begun finish was:

Howland Island → Honolulu → Oakland

The schematic below presents those categories separately. It is a route overview rather than a geographic projection, and the Lae–Howland line represents the intended course—not a recovered aircraft track.

flowchart LR
    OAK["Oakland<br>Physical departure"] --> BUR["Burbank"]
    BUR --> TUC["Tucson"]
    TUC --> NOL["New Orleans"]
    NOL --> MIA["Miami<br>Public world-flight launch"]

    MIA --> AM["Caribbean and<br>northeastern South America"]
    AM --> BRA["Brazil"]
    BRA --> ATL["South Atlantic"]
    ATL --> AFR["West and<br>central Africa"]
    AFR --> ASIA["South and<br>Southeast Asia"]
    ASIA --> IND["Dutch East Indies"]
    IND --> AUS["Northern Australia"]
    AUS --> LAE["Lae<br>Final confirmed departure point"]

    LAE -. "attempted; precise track unknown" .-> HOW["Howland Island<br>not reached"]
    HOW -. "planned; not flown" .-> HON["Honolulu"]
    HON -. "planned; not flown" .-> OAK2["Oakland<br>intended finish"]

    classDef completed fill:#d9ead3,stroke:#38761d,stroke-width:2px;
    classDef attempted fill:#fce5cd,stroke:#e69138,stroke-width:3px;
    classDef unflown fill:#eeeeee,stroke:#666666,stroke-width:2px,stroke-dasharray:5 5;

    class OAK,BUR,TUC,NOL,MIA,AM,BRA,ATL,AFR,ASIA,IND,AUS,LAE completed;
    class HOW attempted;
    class HON,OAK2 unflown;

A finished map needs five visually distinct evidence categories:

Map treatment	Meaning
Solid line	Completed legs between reported landing points
Attempted-course line	The planned Lae–Howland course, which was begun but not completed
Separate dashed line	Wholly unflown Howland–Honolulu–Oakland legs
Point or annotation	Reported information, such as a radio message
Shaded or dotted corridor	Uncertain movement or a disputed disappearance hypothesis

Oakland and Miami require different labels. Oakland was the physical starting point of the repaired second attempt; Miami was where Earhart publicly announced the circumnavigation and began its international phase. Treating both as “the start” without explanation creates a contradiction that the chronology itself resolves.

The completed line should end at Lae. From there, an attempted-course line can point toward Howland, accompanied by radio-report annotations and an explicit uncertainty notice. It must not imply that the Electra followed the ideal course all the way to the island.

Routes toward Nikumaroro, the Marshall Islands, Saipan, or a presumed ocean-crash area are disappearance hypotheses. They belong in shaded or dotted comparison corridors, not in the completed or intended itinerary.

Why the final attempt flew east instead of west

Earhart made two distinct world-flight attempts in 1937. Combining them produces the wrong direction, crew, and route.

The first attempt was westbound from Oakland. Earhart reached Hawaii, but the Electra was badly damaged during a takeoff accident at Luke Field on March 20, 1937. The aircraft was returned to California for repairs, ending that attempt as documented in a historical reconstruction published by the U.S. Naval Institute.

That first attempt also had a different crew arrangement. Fred Noonan and Harry Manning served as navigators, while pilot Paul Mantz accompanied the flight as far as Hawaii. After the aircraft was repaired, Noonan was Earhart’s only crew member on the second attempt.

The repaired flight reversed direction and traveled east. The U.S. Naval Institute account attributes the change partly to the seasonal window for crossing the Caribbean and Africa before average conditions became less favorable later in June and early July. A separate reported account emphasizes the advantage of reaching India before the peak summer monsoon period in The New Yorker’s history of the two attempts.

These explanations support the same broad conclusion: the reversal reflected a revised operating schedule and seasonal weather considerations. It did not change the overall aim of an equatorial or near-equatorial circumnavigation ending back in Oakland.

Feature	March attempt	May–July attempt
Initial direction	Westbound	Eastbound
Opening route	Oakland toward Hawaii	Oakland across the continental United States
Outcome	Ended after the Hawaii takeoff accident	Reached Lae; disappeared during the attempted Howland leg
Navigating crew	Fred Noonan and Harry Manning	Fred Noonan alone
Map status	Separate, abandoned attempt	Final route covered here

The second itinerary was therefore not simply the first route flown backward. Direction, timing, crew composition, and some proposed stops changed. A route map of the final attempt should not mix in locations that belonged only to preliminary proposals or the failed March flight.

Oakland to Miami: where did the world flight actually begin?

The most accurate short answer is Oakland physically, Miami publicly.

Earhart and Noonan left Oakland on May 20, 1937. The Electra then crossed the southern United States through Burbank, Tucson, and New Orleans before reaching Miami. Earhart announced the circumnavigation there and departed for San Juan on June 1.

Departure date	From	To	Route status
May 20	Oakland, California	Burbank, California	Completed
May 21	Burbank	Tucson, Arizona	Completed
May 22	Tucson	New Orleans, Louisiana	Completed
May 23	New Orleans	Miami, Florida	Completed
June 1	Miami	San Juan, Puerto Rico	Completed; public launch of the announced world flight

Table note: These dates and crew details come from a secondary itinerary reconstruction rather than a surviving comprehensive flight log. The compilation also reports that George Palmer Putnam and mechanic R. D. McKneeley accompanied Earhart and Noonan from Oakland to Miami; Earhart and Noonan continued alone from Miami in the compiled route chronology.

Miami appears as the starting point in many summaries because that was where Earhart publicly announced the circumnavigation and began the international sequence. In that public sense, Miami was the launch. It was not, however, the first place from which the repaired Electra moved during the final attempt.

Omitting Oakland would leave out a completed section of the physical journey. Conversely, describing Oakland–Miami as a separate circumnavigation would detach it from the eastbound attempt it positioned the aircraft to undertake.

The clearest map labels are therefore:

• Oakland — physical departure, May 20
• Miami — public world-flight launch, June 1

Across the Americas and Africa: Miami to Assab

From Miami, Earhart and Noonan crossed the Caribbean to Puerto Rico, continued through northeastern South America, followed Brazil’s Atlantic side, crossed the South Atlantic, and traversed Africa toward the Red Sea.

The chronology below is a secondary-source reconstruction. It is useful for plotting the reported sequence, but its precision should not be mistaken for a fully verified primary flight log. City labels also should not be silently converted into modern airport coordinates.

Departure	From	To	Route note
June 1	Miami, Florida	San Juan, Puerto Rico	International phase begins
June 2	San Juan	Caripito, Venezuela	Caribbean to South America
June 3	Caripito	Paramaribo, Suriname, then Dutch Guiana	Northeastern South America
June 4	Paramaribo	Fortaleza, Brazil	Toward Brazil’s Atlantic coast
June 6	Fortaleza	Natal, Brazil	Positioning for the Atlantic crossing
June 7	Natal	Saint-Louis, Senegal, then French West Africa	South Atlantic crossing
June 8	Saint-Louis	Dakar, Senegal	Short regional leg
June 10	Dakar	Gao, Mali, then French West Africa	Inland across West Africa
June 11	Gao	Fort Lamy, now N’Djamena, Chad	Central African leg
June 12	Fort Lamy/N’Djamena	El Fasher/Al Fashir, Sudan	Eastward across Africa
June 13	El Fasher/Al Fashir	Khartoum, Sudan	Arrival at Khartoum
June 13	Khartoum	Massawa, then Italian Eritrea	Continued on the same date
June 14	Massawa	Assab	Red Sea coast
June 15	Assab	Karachi, then British India	Departure from Africa

Table note: The stop sequence, departure dates, and description of Saint-Louis as an unscheduled landing associated with navigational error are drawn from TIGHAR’s secondary leg-by-leg route compilation. A user-created mapped chronology separately reports that the Natal–Saint-Louis crossing took 13 hours and 22 minutes; that duration is a secondary claim, not a direct reading from an authenticated flight log on the Tripline itinerary.

The Saint-Louis landing illustrates the difference between a landing sequence and a preflight plan. The compiled itinerary identifies it as unscheduled and attributes it to navigational error. Earhart and Noonan continued to nearby Dakar the next day.

Historical names also require care. Fort Lamy is now N’Djamena; El Fasher and Al Fashir are alternate renderings; and colonial descriptions in period reports do not correspond neatly to current national labels. Pairing the old and modern names helps readers locate the same stop across historical and present-day maps.

That does not justify false coordinate precision. A named city may establish the reported stop well enough for a general route reconstruction without proving a particular runway threshold or the coordinates of a modern airport. Where contemporary aerodrome records have not been verified, a city-level marker is the more honest choice.

By Assab, the Electra had crossed the Americas, the South Atlantic, and most of Africa. The route then continued through South and Southeast Asia, the Dutch East Indies, and northern Australia.

Across Asia and Australia: Assab to Lae

From Assab, Earhart and Noonan crossed toward Karachi and proceeded east through the Indian subcontinent and Southeast Asia. They then flew through the islands of what was then the Dutch East Indies before reaching Darwin and Lae.

The reported sequence was:

Assab → Karachi → Calcutta → Akyab/Sittwe → Rangoon/Yangon → Bangkok → Singapore → Bandoeng/Bandung → Soerabaja/Surabaya → Bandoeng/Bandung → Koepang/Kupang → Port Darwin/Darwin → Lae

These paired labels do not represent duplicate stops. Akyab is now Sittwe, Rangoon is now Yangon, Bandoeng is the historical spelling of Bandung, Soerabaja corresponds to Surabaya, and Koepang to Kupang. “Port Darwin” and “Darwin” likewise refer to the same destination in this itinerary.

Period	Reported sequence	Qualification
June 15–18	Assab → Karachi → Calcutta → Akyab/Sittwe	Long transition from Africa into South Asia
June 19–21	Akyab/Sittwe → Rangoon/Yangon → Bangkok → Singapore → Bandoeng/Bandung	Weather interrupted progress around Akyab and Rangoon
June 22–27	Bandoeng/Bandung → Soerabaja/Surabaya → Bandoeng/Bandung	Included a layover and operational work
June 27–29	Bandoeng/Bandung → Koepang/Kupang → Port Darwin/Darwin → Lae	Final completed regional sequence

Table note: The dates, historical place names, and stop sequence are taken from a detailed secondary itinerary. That compilation reports that weather forced a return to Akyab during an attempted continuation and also affected movement around Rangoon.

The evidence does not support assigning one cause to every delay. Stops and layovers could involve refueling, crew rest, maintenance, weather avoidance, or operational checks. More than one factor may have applied at a particular location.

Historical analysis reports substantial maintenance at Karachi, Bandoeng, and Lae, illustrating the continuing demands placed on the Electra during the journey in the U.S. Naval Institute’s account of the flight’s planning and support.

Bandoeng appears twice because the aircraft flew from there to Soerabaja and subsequently returned. Joining the two cities only once would erase part of the reported sequence. Treating Bandoeng and Bandung as separate places would create the opposite error.

Earhart and Noonan reached Lae on June 29, 1937, according to the compiled itinerary. They remained there for maintenance and preparations before departing on July 2.

On the route map, Lae should be labeled final confirmed departure point. It was a staging point for the next Pacific leg, not the planned endpoint of the circumnavigation.

Why Howland Island was essential—and so difficult to find

Howland Island was not an optional detour. The Electra’s limited range made it a necessary refueling stop before the planned continuation to Hawaii.

The Lae–Howland leg was approximately 2,556 miles and was expected to take about 18 hours. The same account describes Howland as roughly two miles long and one mile wide—a small, low, isolated target at the end of a long open-ocean flight in the National Air and Space Museum’s overview.

A runway and fuel had been prepared on the island. The Coast Guard cutter Itasca waited nearby to communicate with the Electra and assist its approach. That support plan nevertheless depended on the ship and aircraft receiving and interpreting compatible signals.

Three navigation methods were meant to work together:

Each method had limitations. Dead-reckoning errors grow when actual winds or groundspeed differ from the estimates.

The methods were most valuable when they reinforced one another. But headwinds could reduce groundspeed and increase fuel consumption, clouds could obstruct observations, and radio problems could remove the expected final guidance.

The Smithsonian’s navigation analysis identifies a combination of limited range, headwinds, clouds, inadequate long-range communication preparation, and radio-direction-finding difficulties. It also emphasizes that the exact errors responsible for missing Howland cannot now be established in its educational reconstruction of the navigation failure.

Dependable two-way communication was never established. The Itasca received transmissions from Earhart, but the exchange required to guide the aircraft to the island did not materialize. Neither side obtained the dependable direction-finding result on which the approach plan depended.

The risks could compound:

• A headwind could reduce groundspeed.
• A lower groundspeed could shift the estimated arrival time.
• Uncertainty in arrival time could enlarge the dead-reckoning error.
• Clouds could prevent or reduce the value of a celestial observation.
• An uncertain position could increase dependence on radio guidance.
• Radio difficulties could then leave the crew without a usable bearing.
• Howland’s small, low profile could make visual detection difficult.

This is a description of interacting risks, not proof of one decisive cause. The available evidence does not establish that one particular wind, cloud layer, equipment problem, navigation error, or fuel calculation caused the disappearance.

A map may show an illustrative range envelope, but it should identify the assumptions behind it rather than present a precise maximum-radius circle as a known boundary.

Lae to Howland: the last documented stage of the flight

The secure chronology begins with departure from Lae on July 2, 1937. Earhart and Noonan intended to fly east to Howland Island. They never landed there.

The Electra transmitted messages that the Itasca received, but the crew and cutter did not establish the dependable two-way exchange needed for a guided approach. Reports that the received signal grew stronger were interpreted as consistent with the aircraft approaching the Howland area. Signal strength alone, however, cannot establish an exact position, distance, or bearing.

The final navigational report is commonly rendered as a statement that the aircraft was on the 157°–337° line of position. A further phrase about “running north and south” appears in commonly used accounts, but the surviving secondary transcriptions and their division into separate messages are not fully consistent.

A line of position is a line somewhere along which the navigator estimates the aircraft lies. It is not a latitude-and-longitude fix:

                         337°
                           /
                          /
       possible aircraft / possible aircraft
             position   /      position
                        /
              reported line
                      /
                     /
                    /
                  157°

The reciprocal figures describe the same line in opposite directions: approximately northwest toward 337° and southeast toward 157°. They do not identify where on that line the aircraft was.

Nor does the report prove that the calculated line passed precisely through Howland. Modern diagrams often center the line on the island to illustrate the intended navigation geometry. That is a mapping convention, not a recovered aircraft position.

The wording about “running north and south” should not be silently converted into a confirmed movement along the 157°–337° line. That line is oriented northwest–southeast, not literally north–south. The phrase may describe a search, an imprecise directional statement, a transcription problem, or a message whose relationship to the line report is uncertain.

Published versions illustrate the difficulty. The National Air and Space Museum page reproduces an anomalous “156–137” form, while other secondary accounts use the reciprocal 157°–337° wording. A personal geographic analysis also centers the line on Howland only as an illustrative assumption and acknowledges uncertainty in interpreting it in its comparison of locations and theories.

The final-stage evidence is best visualized as an uncertainty diagram rather than a continuous track:

flowchart LR
    LAE["Lae<br>departed July 2"] -. "planned course; actual track unknown" .-> HOW["Howland Island<br>intended destination"]

    RADIO["Received radio reports<br>do not provide coordinates"] --> LOP["Reported 157°–337°<br>line of position"]
    LOP --> UNKNOWN["Final position and track<br>unknown"]

    UNKNOWN -. "hypothesis only" .-> SEA["Ocean-crash area<br>not established"]
    UNKNOWN -. "hypothesis only" .-> NIK["Nikumaroro"]
    UNKNOWN -. "less-supported hypothesis" .-> MAR["Marshall Islands"]
    UNKNOWN -. "less-supported hypothesis" .-> SAI["Saipan-linked scenario"]

    classDef secure fill:#d9ead3,stroke:#38761d,stroke-width:2px;
    classDef attempted fill:#fce5cd,stroke:#e69138,stroke-width:3px;
    classDef report fill:#d9eaf7,stroke:#3d85c6,stroke-width:2px;
    classDef unknown fill:#eeeeee,stroke:#333333,stroke-width:3px;
    classDef hypothesis fill:#eadcf8,stroke:#674ea7,stroke-width:2px,stroke-dasharray:5 5;

    class LAE secure;
    class HOW attempted;
    class RADIO,LOP report;
    class UNKNOWN unknown;
    class SEA,NIK,MAR,SAI hypothesis;

Any reconstruction of the exact route faces serious limits. Actual winds aloft are not continuously known; forecasts and ship observations are not a record of what the Electra experienced at every point. Cloud cover, groundspeed, fuel state, and the search pattern also remain uncertain. One aviation research project explicitly labels its own proposed Lae–Howland reconstruction a hypothesis because essential operational and meteorological data are missing in its discussion of the final flight.

The responsible record ends with these points:

• Earhart and Noonan left Lae for Howland.
• The Itasca received transmissions from the Electra.
• Dependable two-way communication and radio direction finding were not established.
• Signal-strength reports were interpreted as consistent with an approach toward the Howland area but did not locate the aircraft.
• A 157°–337° line of position was reported in commonly used transcriptions.
• The relationship between that line and the disputed “north and south” wording is not settled.
• The precise track and location after the final reliable reports remain unknown.

The unfinished route, mileage differences, and disappearance theories

Had Earhart landed on Howland, the intended final sequence was:

Howland Island → Honolulu → Oakland

Neither leg was flown. A secondary route compilation lists approximately 1,900 miles from Howland to Honolulu and 2,410 miles from Honolulu to Oakland. It reports 28,595 miles for the full planned itinerary and estimates 24,285 miles traveled from Oakland toward the vicinity of Howland in its mileage table for the world-flight route.

Those numbers should not be combined with other totals as if they all measure the same thing. The user-created mapped itinerary displays 24,211 miles, while broader descriptions often round the intended journey to roughly 28,000 or 29,000 miles.

Mileage label	What it may measure
Planned itinerary	Every intended leg through the return to Oakland
Completed landings	Legs ending at reported places the aircraft reached
Mapped route	Distances calculated between a particular set of map points
Estimated progress toward Howland	An assumed portion of the final uncompleted leg
Rounded public total	A simplified description of the overall plan

No single total can accurately represent both the exact completed distance and the full planned circumnavigation. Earhart did not land at Howland, yet some mileage estimates include presumed progress toward its vicinity.

The same distinction applies to disappearance theories. They propose what might have happened after Earhart failed to find Howland; they do not extend the confirmed flight path.

Proposed outcome or route	Relationship to reported line	Fuel assumption	Evidence commonly cited	Major uncertainty
Fuel exhaustion followed by an ocean crash	Compatible with an unsuccessful search near the intended destination	Assumes usable fuel was eventually exhausted before another landing	Failure to reach Howland, radio and navigation difficulties, limited range, unsuccessful search	No wreckage or exact impact point has been established
Southward continuation to Nikumaroro	Geometrically compatible with movement toward the southeast side of the reported line	Requires enough fuel to continue from the search area	Claimed post-loss signals, artifacts, a historical photograph, and later analysis of recorded bone measurements	Evidence is circumstantial, disputed, and not uniquely linked to Earhart
Northwest continuation toward the Marshall Islands	Could be drawn beyond the northwest side of the line, but no track confirms it	Requires sufficient endurance and a route far from Howland	Later testimony and capture narratives	Weak connection to the documented navigation record
Saipan-linked travel or capture	A direct route would require a major departure from the intended course	Requires a large deviation, substantial endurance, or an intermediate event	Oral histories and capture claims	No confirmed route, landing, or documentary chain establishes the scenario

Fuel exhaustion followed by an ocean crash is a leading institutional assessment, not proof of a particular crash point. The Smithsonian concludes that Earhart and Noonan never reached Howland and likely died in the ocean, while expressly acknowledging that the precise errors and final location may never be known in its navigation assessment.

The Nikumaroro proposal is more specific. TIGHAR argues that the Electra may have continued southeast along the reported line toward Gardner Island, now Nikumaroro, landed on the reef, and left Earhart and Noonan there as castaways. Its case invokes reported post-loss radio signals, artifacts, historical imagery, and later analysis of measurements from bones found on the island.

Those claims remain contested. The bones themselves were lost; surviving analysis depends on recorded measurements. Artifact associations are disputed, reported post-loss signals are not conclusively authenticated, and no cited item uniquely identifies Earhart, Noonan, or the Electra. The route and survival narrative therefore remains TIGHAR’s reconstruction rather than an established extension of the flight path as presented in an aviation-magazine interview about the hypothesis.

Their locations can appear on a comparative theory map, but only as labeled hypothesis areas—not as completed, attempted, or planned route segments.

The final map must leave readers with three separate statements:

Intended destination: Howland Island, followed by Honolulu and Oakland.

Reported navigation information: A 157°–337° line of position, accompanied in some transcriptions by disputed wording about running north and south.

Ultimate location: Unknown.

Earhart’s journey can be traced from Oakland to Lae, and her destination after Lae is clear. What cannot be drawn as fact is the Electra’s final movement after it failed to find Howland.

Did Amelia Earhart’s final world flight start in Oakland or Miami?

Both, in different senses. The physical journey began in Oakland on May 20, 1937. The Electra then traveled through Burbank, Tucson, and New Orleans before reaching Miami. Earhart publicly announced the circumnavigation there and departed for San Juan on June 1.

A map should label Oakland “physical departure” and Miami “public world-flight launch.” The Oakland–Miami section was part of the final eastbound attempt, not a separate circumnavigation.

How far was the flight from Lae to Howland Island?

The planned Lae–Howland leg was approximately 2,556 miles and was expected to take about 18 hours according to the National Air and Space Museum’s account of the final stage.

That figure describes the planned leg, not a measured final track. Winds, deviations, course corrections, and search maneuvering could have changed the distance actually flown.

What does the 157°–337° line of position mean?

It means the navigator estimated that the aircraft was somewhere along a line oriented approximately northwest to southeast. The figures are reciprocal directions: 157° one way and 337° the other.

A line of position is not a unique latitude-and-longitude fix. Earhart and Noonan could have been anywhere along the calculated line, and the evidence does not prove that it passed directly through Howland.

The phrase about “running north and south” should be treated separately. Secondary transcriptions differ, and literal north–south movement is not the same as movement along a 157°–337° line.

Where was Amelia Earhart when she disappeared?

She was flying over the Pacific from Lae toward Howland Island, but her exact position is unknown. The Itasca received radio transmissions as the Electra approached the expected Howland area, yet signal strength and the reported line of position cannot provide a precise coordinate.

No verified evidence establishes whether the aircraft was north, south, east, or west of Howland at the final reliable transmission. Nor does it establish a crash point or subsequent landing place.

Was Nikumaroro part of Amelia Earhart’s confirmed flight path?

No. Nikumaroro—then called Gardner Island—was not a scheduled stop and is not part of the completed or intended itinerary.

It appears in TIGHAR’s disputed hypothesis that Earhart and Noonan continued southeast after failing to find Howland. The proposal relies on circumstantial and contested radio, artifact, photographic, and bone-measurement claims. Nikumaroro should therefore be marked only as a hypothesized destination, never as a confirmed flight-path point.