For decades, the Bermuda Triangle has been associated with stories of ships and aircraft disappearing without a trace. A 2001 sonar survey off the coast of Cuba produced images of large, geometrically regular blocks on the ocean floor at a depth of about 700 meters. No follow-up expedition has returned to the site, and no scientific consensus has been reached about what the images show.
The name “Bermuda Triangle” is more recent than many of the incidents associated with it. Writer Vincent Gaddis coined the term in a 1964 magazine article, decades after several famous disappearances in the region. That timing matters because the mystery was assembled retroactively, grouping together incidents that spanned half a century before anyone compared them systematically.
One of the earliest cases involved the USS Cyclops, a U.S. Navy ship that left Barbados on March 4, 1918, with 309 men aboard and a cargo of manganese ore bound for Baltimore. The weather was moderate, the ship was in good mechanical condition, and its radio was working. The ship disappeared along the route without sending a distress signal, and no wreckage was ever found. The Navy conducted one of its largest searches in history and found nothing.
Two sister ships, the Proteus and the Nereus, nearly identical in design and carrying similar heavy metal cargo, vanished along roughly the same Atlantic corridor in 1941. Navy investigators suggested the three ships may have suffered structural failures from overloading, but that theory was never confirmed. The Cyclops loss remains the largest non-combat loss of life in U.S. Navy history.
In December 1945, five U.S. Navy Avenger torpedo bombers departed Fort Lauderdale on a routine training mission designated Flight 19. The weather was clear. Within two hours, the flight leader, Lieutenant Charles Taylor, reported confusion over the radio. His compass was behaving oddly, and he could not determine his location. Communications broke down and then stopped. A PBM Mariner rescue plane, a type known for documented fuel vapor explosions in its enclosed bomb bay, was dispatched and also disappeared with its crew of 13. No wreckage from any of the six aircraft was found.
Various explanations have been proposed over the years, including magnetic anomalies, sudden storms, and rogue waves, which oceanographers have since confirmed occur in the region. None have satisfied researchers who studied the individual cases closely. One fact remained largely ignored for decades: the ocean floor beneath the triangle was almost completely uncharted. The surface and the atmosphere above it had been studied, but the bottom remained largely unknown. Today, NOAA estimates that more than 80 percent of the world’s ocean floor still lacks detailed mapping. In the early 2000s, that figure was even higher.
That changed in 2001, when a commercial survey operation sent equipment to the seafloor and found something unexpected. The expedition was not motivated by scientific curiosity. In 2000, Advanced Digital Communications, a Canadian marine engineering company co-founded and led by oceanographer Pauline Zalitzki with her partner Paul Weinstein, signed an agreement with the Cuban government. The arrangement was purely economic. Cuba needed revenue, and the waters around the island have a long maritime history. Hundreds of Spanish ships sank there over the centuries due to hurricanes, naval battles, and the hazards of colonial trade routes, many carrying silver, gold, and other goods that never reached European ports. Spain’s colonial fleet, known as the Flota de Indias, had sailed annual convoys through these waters since the 1560s.
The agreement was straightforward. Zalitzki’s company would use advanced sonar imaging to survey the seafloor, locate shipwrecks, and help recover valuable artifacts. In return, the Cuban government would receive a share of anything found. Cuba had made similar arrangements with other foreign survey groups during the 1990s after the collapse of Soviet support in 1991.
The team consisted of engineers, marine surveyors, and ocean specialists. Zalitzki was an experienced deep-sea surveyor with years of commercial work. Their task was to locate shipwrecks, estimate their contents, and prepare recovery files. The research vessel, later known as the Ulysses, departed with calibrated sonar arrays. The crew fell into the routine of deep-sea survey work, running overlapping grid lines across a target area off Cuba’s western coast near the Guanahacabibes Peninsula.
For days, operators watched screens filled with faint green reflections: soft mud, sedimentary ridges, scattered rock fragments, the usual chaotic terrain of the seafloor. Then an anomalous signal appeared. The first operator logged it as likely equipment noise, a common occurrence with sonar systems. The ship completed its pass and returned to the next grid line. The anomaly remained. Same location, same dimensions, same pattern. A colleague was called over. A pass from a different angle showed the same shape. The team adjusted transmission frequencies, changed the tow depth of the array, and modified signal sensitivity. Nothing changed the image.
After six overlapping passes across the site, the noise hypothesis was no longer viable. Sonar works by sending sound pulses downward and measuring their return to a sensor array. Different materials reflect sound differently. Soft sediment absorbs much of the pulse and returns a weak, dark signal. Dense rock reflects sound strongly and returns brightly. The result is an acoustic image of the seafloor. Normal deep-sea floor appears chaotic on a sonar screen, with randomly scattered rocks and irregular sediment fields. The boundary between materials does not follow a straight line.
The bright acoustic signals from the Cuban site indicated dense, solid material forming long, straight lines. These lines maintained regular spacing, met at various angles, and repeated across an area too large to be explained by a single piece of debris. Operators then noticed acoustic shadows, directional shadows following the bright shapes in the opposite direction of the transmitted pulse. Such shadows only exist when an object rises above the seafloor and blocks the returning sound wave. Flat sand cannot cast acoustic shadows, and scattered debris cannot produce them with this regularity. Whatever was there had height and mass, and it was stable on the seafloor.
The team compared the shapes against reference libraries of known shipwrecks, coral formations, and volcanic outcroppings. Shipwrecks scatter in arcs of structural collapse. Coral grows in branching organic curves. Volcanic rock is compact and irregular. None matched. The seafloor displayed order, and order in a chaotic environment required a direct response. The team decided to send a remotely operated vehicle down to the site.
The ROV descended to 700 meters. Before deployment, the crew checked every system: cameras, LED arrays, thruster response, cable integrity, signal transmission, emergency buoyancy. At that depth, there is no room for repairs. The vehicle entered the water in the early morning. During the first 100 meters, surface light still reached the lens. At 200 meters, it disappeared entirely. Below that, the only light came from the vehicle’s LED array, projecting a cone of white light roughly 15 meters wide.
Pressure at 700 meters is about 70 atmospheres, enough to crush an unprotected steel cylinder. Deep-water ROVs are built with titanium or specially reinforced structures, a technology that only became commercially reliable in the 1990s. Water temperature hovers just above freezing, between 2 and 4 degrees Celsius. Fine sediment, organic material, and mineral dust drift continuously downward, a slow biological rain that oceanographers call marine snow. This is the mechanism that buries almost everything on the deep seafloor over time.
The ROV descended for about 40 minutes. The crew spoke little. Then, at the edge of the light, the seafloor appeared. The first images were difficult to read. Flat forms emerged gradually from the darkness. The operators expected the usual irregular ridges and scattered debris. What the camera showed did not match that expectation. Surfaces were wide and flat where the seafloor should have been uneven. Edges were distinct, not the rounded lines formed in geological materials after thousands of years of pressure and water movement, but true geometry. Angles met at right angles. Flat surfaces extended for continuous meters.
Four-minute descent footage showed massive stone blocks, not one or two but dozens visible across the full range of the camera. Some measured several meters along their longest dimension. Their sides were flat, their surfaces intact. Many were arranged in straight lines with even spacing, a regularity that random geological deposition does not produce. The material itself was not the soft limestone that forms most of the geological layer around Cuba, a rock known to erode into holes, cracks, and rounded surfaces when exposed to seawater and pressure. These blocks were dense and hard, acoustically closer to granite, an igneous rock that is highly resistant to erosion and difficult to cut or shape without special tools.
One detail stopped the team: the block surfaces were exposed. At 700 meters, fine sediment continuously falls and accumulates across the seafloor over geological time, thick enough to bury exposed objects entirely, softening edges and hiding surfaces. These blocks were not buried. Their edges were sharp and clearly defined. Their surfaces were exposed. The top sediment layer was far thinner than the expected accumulation for objects resting on the seafloor for a long geological period. Two possibilities remained: either these objects were much younger than their depth suggested, or a rapid, catastrophic event had deposited them so suddenly that natural sediment accumulation had not yet begun.
The team returned to the surface and spent days reviewing every frame recorded by the ROV cameras. Measurements were estimated from the footage using known reference points on the vehicle. Material density assessments were collected. Sediment coverage was documented. All of this information was sent to independent specialists, including geologists, oceanographers, and marine archaeologists with no prior connection to the survey.
One of those who commented publicly was Manuel Iturralde Vinent, a prominent geologist at Cuba’s National Museum of Natural History, whose early skepticism became influential. Iturralde Vinent argued that regular-looking formations can arise from natural stratification and mineral crystallization without any human intervention, and noted that no formal peer-reviewed geological survey of the actual site had been published. His caution reflected a legitimate scientific standard. Unusual claims about human origin require direct physical sampling, not just sonar interpretation. No core samples or dated sediments have been extracted from the site to settle the question.
Independent reviewers who examined Zalitzki’s footage did not agree on a natural explanation that accounted for all the facts at once. Each proposed geological hypothesis solved part of the puzzle while leaving other parts unresolved. Columnar basalt, formed when cooled lava contracts into geometric columns, was examined and ruled out. The geometry did not fit basalt, and the acoustic signature did not match a volcanic origin. Tectonic fracturing patterns were considered. Large-scale seismic fracturing can produce regular structures in some documented cases, but not at this size, with this degree of surface consistency and even spacing. Submarine volcanic activity was proposed and rejected for similar reasons.
The team expanded the sonar survey grid beyond the initial discovery. The anomaly continued. The more the grid widened, the more organized shapes appeared in the returning signals. The initial group was part of a much larger site. Expanded data showed what appeared to be long flat pathways extending between raised formations, some producing acoustic characteristics consistent with elevated structures with sloping sides. The estimated total area covered roughly the size of a small town. A single group of unusual blocks could be dismissed as coincidence. A site spanning this much survey area, with internal consistency repeating across its full extent, is not something random geological processes typically produce. The footage showed something that looked, practically, like construction.
If the structures were built by human hands, the depth at which they sit determines a timeline. The Caribbean region does not sink 700 meters over a short geological period through ordinary processes. That much subsidence requires a major geological event. Researchers working with regional subsidence rate data produced two competing estimates for when any surface at this location was last above water. The conservative estimate, based on gradual tectonic subsidence models, suggests the surface was above sea level about 50,000 years ago. A second estimate places the site as possibly available during the last ice age peak, which ended about 12,000 years ago, when global sea levels were 120 to 130 meters lower than today.
Even the more recent estimate requires the surface to have been underwater for 12,000 years. The older estimate places it above water at a time when modern humans had not yet reached the Caribbean. Both timelines conflict with the same archaeological framework. The earliest confirmed examples of large-scale stone construction in the world are Göbekli Tepe in modern Turkey, dating to about 9600 BCE, and the megalithic temples of Malta, between 10,000 and 12,000 years old. In the Caribbean specifically, the earliest confirmed human presence dates to roughly 7,000 to 8,000 years ago, associated with the Guanahatabey people, mobile hunter-gatherers documented through simple stone tool assemblages at sites such as Funche Cave in western Cuba.
The shapes on the seafloor suggest a population with no known capability, based on the complete archaeological record, to cut dense granite blocks to precise dimensions and organize them into a city-sized site. The shapes suggest planning, precision, and organizational capacity that, according to current understanding of the region at that time, should not have existed.
One geological mechanism offers a partial explanation: a catastrophic seismic event. A large earthquake with enough force to rapidly displace a section of coastal land into deep water could explain the depth without requiring an extraordinary timespan. Cuba’s coast lies near the active Oriente fault, part of the same tectonic system responsible for Haiti’s devastating 2010 earthquake, and significant historical seismic activity has been recorded. A rapid collapse would also explain the sedimentary evidence. A site that fell suddenly into deep water would be sealed under the water column before ordinary sediment burial began, preserving the sharp edges documented by the ROV. This explains how the site reached the bottom. It does not explain what was built there, who built it, or how.
Zalitzki returned from the 2001 survey with acoustic data, documented ROV footage, and a discovery whose significance she understood. National Geographic and the BBC covered the event. Specialists in deep-sea geology, marine archaeology, and ancient Caribbean history reviewed the material. Many independent reviews agreed that the evidence was genuine, the footage authentic, and no compelling natural explanation existed.
The project then stopped for political and financial reasons. The site lies in Cuban territorial waters, and decades of the U.S. trade embargo, tightened by the Helms-Burton Act of 1996, made large-scale American participation in a Cuban water expedition legally risky. Most institutions capable of deep-sea work, such as Woods Hole, NOAA, and university marine science programs, are American. The diplomatic opening under Obama from 2014 to 2015 came too late and ended too quickly, before any follow-up expedition could be organized.
Cost compounded the problem. Deploying one ROV to 700 meters requires a purpose-built ship, a specialist crew, and armored equipment, while a full systematic survey costs millions of dollars. No organization has secured both funding and access to the site since 2001. There has also been quiet resistance. Whether the formations are natural or artificial, one of the two possibilities means someone with relevant expertise would be wrong. Either the geology is anomalous with no known equivalent, or the accepted timeline of human civilization in the Americas needs revision.
No follow-up has occurred. No samples have been extracted. No blocks have been measured. No submarine has returned. The site remains as it was left in 2001. Answering the question requires three things that have not yet come together: access to the site through Cuba, a funded expedition, and an institution prepared to support the results. None of that is impossible. None of it has happened. The blocks remain at 700 meters in total darkness, awaiting another look.
