One sonar pass. Twenty-five years of silence. The file is still open.

// The Survey

Advanced Digital Communications. Victoria, British Columbia. Licensed by the Cuban government as one of four joint-venture firms cleared to survey Cuban territorial waters. Commercial brief: shipwrecks and oil.

The principals were Paulina Zelitsky — Odesa Marine Engineering 1967, University of Alberta 1976, P.Eng. Canada — and her husband Paul Weinzweig. Zelitsky had defected from the Soviet Cuban submarine-base programme in 1971. The deep-water surveying skill set ADC brought to its Cuban joint venture was the same Cold-War skill set she had been trained in by the side that had lost the contract.

The Cuban scientific lead was Dr. Manuel A. Iturralde-Vinent of the Cuban Academy of Sciences — former Deputy Director of the National Museum of Natural History in Havana, later (2007–2016) President of the Cuban Geological Society, a working palaeontologist with a multi-decade publication record across the Caribbean stratigraphic literature. He was not an external reviewer dropped onto the project after the fact. He was inside it from the dating runs.

The first relevant pass came off R/V Ulises in July 2000. Side-scan sonar in a standard commercial salvage configuration. The bottom feature the processing operator flagged on the return — a roughly two-square-kilometre footprint in the south-eastern San Antonio Lowlands, coordinates 21°46′21″N 84°50′12″W, approximately thirteen kilometres off the tip of Cabo de San Antonio — showed symmetrical lithic forms two to five metres across, arrayed in perpendicular and circular patterns. Iturralde-Vinent's own 2021 description: “cubic, parallelepiped and pyramidal blocks with relatively polished faces.”

ADC returned in 2001 with a 1.3-tonne unmanned remotely-operated vehicle on fibre-optic tether, video cameras and lights at 600–750 metres. The ROV passed over the formation. ADC named the site MEGA — the label Iturralde-Vinent has used in his subsequent scientific synthesis.

A third survey followed in spring 2002 for sample acquisition: magnetometer survey, side-scan, ROV-recovered mineral and biological specimens. Iturralde-Vinent's geological team participated directly. The samples have never been published in English-language peer-reviewed literature.

// The Public Record, 2001–2002

Reuters broke the story 14 May 2001 in a Havana wire by Andrew Cawthorne, headlined Explorers Comb Cuban Seas for Treasure, Mysteries. The story carried two voices — ADC's claim of “extraordinary stone formations,” and Iturralde-Vinent's caution.

BBC News followed on 7 December 2001 under the headline ‘Lost city’ found beneath Cuban waters. The article carried Iturralde-Vinent's now-most-quoted line: “These are extremely peculiar structures, and they have captured our imagination. But if I had to explain this geologically, I would have a hard time.”

National Geographic News ran Brian Handwerk's New Underwater Finds Raise Questions About Flood Myths on 28 May 2002. The piece situated the site against the Younger Dryas hypothesis and quoted Sylvia Earle — the National Geographic Society's resident oceanographer, then heading the Sustainable Seas Expeditions programme — describing the site as “compelling… we need to go check it out.” A National Geographic manned-submersible expedition was tentatively scheduled for later in 2002 under the Sustainable Seas umbrella.

Two months later, the Washington Post Foreign Service ran Kevin Sullivan's In Cuban Depths, Atlantis or Anomaly? on 10 October 2002, which thickened Iturralde-Vinent's position: “It's strange, it's weird; we've never seen something like this before.” The same piece carried Robert Ballard, the oceanographer who located the Titanic, deflating the imagery: “That's too deep. I'd be surprised if it was human… sonar images can be sort of like looking at an ink blot — people can sometimes see what they want to see.”

The public record at peak: a salvage firm with a sonar pass and a 1.3-tonne ROV, a Cuban Academy of Sciences geologist who could not explain the formation geologically and would not commit to a non-natural interpretation, an oceanographic survey programme at the National Geographic Society proposing to take a manned submersible to depth, and one of the most credible voices in deep-ocean exploration counselling against the imagery.

That was the peak. The descent followed inside twelve months.

// The Geological Counter-Read

The principal sceptical voice across the 2001–2002 cycle, and the only voice that ever published a scientific synthesis on the site, was Manuel Iturralde-Vinent.

The harshest read of the imagery came from the geologist who was inside the project. He was not an external reviewer drafted in to debunk. He was the Cuban Academy of Sciences geological lead embedded in ADC's own scientific advisory from the first dating run. The Reuters and BBC quotes were not the verdict of an outside critic. They were the holding pattern of an embedded scientist who refused to commit either way without samples in hand.

By 2021 he had committed. Estructuras líticas submarinas al SW de Cuba (“Submarine lithic structures off SW Cuba”), posted to ResearchGate in June 2021, is the only known scientific synthesis of the MEGA site by a working geologist with direct access to the dataset (Iturralde-Vinent 2021 ↗).

The specific anomalies the 2001–2002 reporting flagged — well-rounded pebbles inside the formation, volcanic scoria fragments where no nearby volcanism exists — are read in the 2021 paper as evidence of transport, not construction. Pre-existing sub-aerial materials swept downslope into the San Antonio depression by submarine soil flow at the geomorphologically appropriate moment.

The depth–dating arithmetic compounds the geological case. The formation sits at 600–750 metres. For human construction to be the explanation, the seafloor under it would have had to subside roughly that distance within human prehistory — fifty thousand years at the conservative end of the subsidence estimate for the local lithology. No archaeological framework currently in print can accommodate intentional construction at that timescale. The Florida State University underwater-archaeology programme's voice, carried in the 2002 Washington Post piece, was the discipline's working summary: “The structures are out of time and out of place.”

The geological counter-read is therefore not a debunking. It is the inside-the-project assessment that the natural mechanism is the simpler explanation of what is documented, and that the unusual features of the documentation can be accommodated by the same mechanism. Iturralde-Vinent's 2021 paper labels its conclusions preliminary.

// The Silence

The 2002 National Geographic Sustainable Seas expedition was cancelled before it sailed.

The documented reason is funding. The Sustainable Seas programme operated on Goldman Fund grants and the National Geographic Society's own line items. A budget constraint that summer collapsed the deployment window. The expedition was rescheduled tentatively for 2003 and never executed at the original scale.

The conspiracy-flavoured version of this collapse — a two-million-dollar National Geographic commitment withdrawn under US State Department pressure — is not in any primary-source coverage. It traces to ancient-civilisation secondary writers in the years after 2002 and has been repeated since by venues that do not cite their sources. The Washington Post (Sullivan, October 2002) attributed the cancellation to budget. Subsequent commentary by Earle and her colleagues, where it exists, has not contradicted that.

A reduced minisub return in 2003 has been mentioned in Zelitsky's later interviews. No peer-reviewed publication followed. No widely circulated NG-branded film. No museum exhibition. The story did not so much get debunked as fail to acquire the institutional follow-up it would have needed to either close or open the file.

From 2003 to 2020 the scientific literature is silent on the MEGA site. The popular literature continued — Atlantis-adjacent venues recycled the 2001–2002 imagery and quotes through their own framing. The file was not contested by anyone with direct access to the data, because no one with direct access to the data was publishing.

Iturralde-Vinent's June 2021 ResearchGate paper broke the silence. It is the first and so far only attempt by a geologist with access to the dataset to interpret it in writing. Its register is preliminary, its conclusion is natural-mechanism, its method is constrained by the same access limitation that constrains everyone — Iturralde-Vinent's interpretation rests on the same dataset ADC collected in 2001 and 2002, not on a re-survey.

The MEGA site has not been independently re-acquired by a third party at the published imagery resolution. To the best of the public record, the dataset that exists is the dataset ADC collected, the firm that collected it is no longer operating at scale in Cuban waters, and the political environment for US–Cuban scientific cooperation in the years since has not provided the conditions for a return.

// What the File Cannot Close

Five items in the public record remain open at filing.

One. The complete ADC dataset — sonar mosaics, ROV video at original resolution, sample analyses — has never been deposited in a public scientific archive. It exists, per the Morien Institute interview of June 2002 and Iturralde-Vinent's 2021 references. It is not accessible to third-party review.

Two. The 2001 and 2002 ROV imagery has not been independently re-acquired. Iturralde-Vinent's 2021 synthesis is an interpretation of the original ADC dataset, not a re-survey. The minimal standard for closing a contested formation — independent re-acquisition at the original resolution — has not been met.

Three. The volcanic scoria anomaly is not fully resolved in the 2021 paper. Iturralde-Vinent treats it as evidence of long-distance transport by submarine soil flow. Whether that mechanism is mechanically plausible at the inferred slope angles is a legitimate geomorphological question that the paper does not work through in depth.

Four. The fifty-thousand-year subsidence arithmetic depends on a subsidence model that has not been calibrated against modern bathymetric and tectonic data for the San Antonio Lowlands. The figure is widely cited; the model behind it is not in the public literature.

Five. The Cuban Academy of Sciences has not formally closed the file. Iturralde-Vinent's 2021 paper presents itself as a synthesis of preliminary results.

The file is open by the geologist's own framing.