The Bermuda Triangle

Ships and aircraft have been lost in the area, as in every busy sea. The western North Atlantic carries heavy traffic to and from US East Coast and Gulf ports, the Bahamas and the Caribbean, and private boating and flying from Florida. Most Atlantic hurricanes cross it; the Gulf Stream can raise sudden, violent seas; and the many islands make for shallow, hazardous water⁠[1].

The best-documented loss is Flight 19. On 5 December 1945 five Navy TBM Avengers with 14 men left Naval Air Station Fort Lauderdale at 2:10 pm on a training flight, Navigation Problem No. 1, which ran east to bomb a target near Bimini, north over Grand Bahama and back⁠[9]. At about 3:40 pm the flight leader, Lt Charles C. Taylor, radioed that both his compasses were out and that he believed he was over the Florida Keys; he was in fact north-east of Fort Lauderdale. The flight changed course repeatedly as the pilots disagreed about where they were. The last transmission, at 7:04 pm, said they would ditch together when the first aircraft fell below ten gallons of fuel; the sea was rough and night had fallen⁠[9]. A PBM-5 Mariner flying boat with 13 men, sent from Banana River to search, was lost the same evening; the tanker SS Gaines Mills reported a burst of flames at 7:50 pm and the escort carrier USS Solomons lost the aircraft on radar at the same position⁠[10].

For the individual cases the search has been in archives. Kusche's method was to set each story in the popular books against what was recorded at the time; the solution, he said, came from interlibrary loan, not from an expedition⁠[4]. For Flight 19 the Navy's own investigation records survive at the National Archives at Fort Worth, including the map of the navigation problem and the correspondence⁠[9]. The Navy board concluded that the flight leader had lost his position; after his mother's appeal the Board for Correction of Naval Records amended the finding to cause unknown⁠[10]. The amendment changed the attribution of blame, not the account of what happened: no wreckage of either aircraft has been recovered⁠[9].

For the pattern as a whole the search is statistical. NOAA's conclusion rests on comparison with other busy seas⁠[1]. Insurers' casualty data give the best modern counts: of 905 total losses of vessels over 100 gross tons worldwide from 2016 to 2025, 21 were in the West Indies region, against 255 in the waters from the South China Sea to the Philippines and 120 in the East Mediterranean and Black Sea⁠[7]. A 2013 WWF study of shipping casualties since 1999, also built on Lloyd's List Intelligence data, found the same hotspots and did not place the Caribbean or the Triangle among them⁠[8]⁠[11]. Neither study divides losses by traffic, and the insurers' West Indies region is not the Triangle, so they show that the area is unremarkable in raw counts, not that its rate per voyage is ordinary.

The ordinary explanation is that the losses are what weather, currents, traffic and human error produce in any busy sea, and that the pattern is an artefact of selective retelling. Kusche's findings support both halves: incidents with ordinary causes, and a list padded with cases that happened elsewhere, at other times, or not as told⁠[4]⁠[5]. NOAA and the Coast Guard take the same view⁠[1]⁠[6].

Two physical mechanisms have been proposed to explain losses without trace. Methane erupting from seabed sediments could in principle reduce buoyancy: laboratory and simulation work shows that a single large bubble can sink a floating body of similar size, though the study was motivated by the North Sea and does not estimate how often such events occur anywhere⁠[12]. NOAA lists methane among the more grounded ideas "if not in evidence"⁠[1]. Compass behaviour is the second: NOAA notes evidence that in parts of the area a magnetic compass can point to true rather than magnetic north, a navigational hazard rather than an anomaly⁠[1].

The anomalous explanations (lost Atlantis, vortices, extraterrestrials, electromagnetic forces) are the ones Berlitz canvassed⁠[3]. They rest on the premise of an excess of unexplained losses, which the case-level research removed⁠[4].

It is settled at the level of cases and official assessment. A quantitative close would be a published loss rate per unit of traffic for a fixed Triangle polygon and for comparison seas of similar traffic and weather, with confidence intervals. The pieces exist in open data: charted wrecks from NOAA's electronic charts, civil aviation accidents from the NTSB database, vessel traffic from AIS, and hurricane tracks from IBTrACS⁠[13]⁠[14]⁠[15]⁠[16]. Lloyd's List Intelligence casualty data, through an insurer or WWF, would allow the same comparison for merchant ships worldwide⁠[7].