The Eye of the Sahara: Unpacking the Atlantis Claims at the Richat Structure

In 1964, three French researchers published a short note in the Comptes Rendus of the Académie des Sciences reporting that they had found coesite in the rocks of the Richat Structure in Mauritania. Coesite is a high-pressure form of silica. On Earth’s surface it forms in essentially one circumstance: the shock of a large meteorite strike. That single mineral identification turned a circular landform in the Sahara into a probable impact crater, and for five years the scientific literature treated it as one. What happened next is a better story than any of the claims later built on this place, and it is fully documented.

A Bullseye in the Adrar

The Richat Structure sits on the Adrar plateau of Mauritania, in the northwestern corner of the Taoudenni Basin, near the caravan town of Ouadane. It is a slightly elliptical depression roughly forty kilometres across, ringed by concentric ridges that are clearly visible from orbit and were first noticed on aerial reconnaissance photographs in the 1930s.

The rings are not walls. Matton, Jébrak and Lee, writing in the journal Geology in 2005, describe them as circular cuestas: three nested ridges that dip outward from the centre, formed where erosion stripped a domed sequence of rock and left the harder layers standing proud. The sequence itself runs from the Late Proterozoic to the Ordovician, hundreds of millions of years of sandstone, quartzite, limestone and dolomite laid down long before anything walked on them.

At the centre the same paper describes a limestone-dolomite shelf enclosing a siliceous breccia measured in kilometres, cut through by basaltic ring dikes, kimberlitic intrusions and alkaline volcanic rocks. That combination is unusual, which is much of why the structure has drawn geologists for seventy years. It is also entirely ordinary in kind: every component is a rock type known from other places, arranged by processes that operate everywhere.

One number is worth holding on to for what follows. The structure stands roughly four hundred metres above sea level, several hundred kilometres from the Atlantic coast.

The Coesite That Wasn’t

The 1964 report of coesite appeared in volume 258 of the Comptes Rendus, at pages 5488 to 5490. It was a short paper making a large claim, and the claim was reasonable on its face: circular structure, plus high-pressure silica, equals impact.

Robert Fudali went back to the samples. In 1969 he published the result in Science, volume 166, issue 3902, pages 228 to 230, under a title that gives away the conclusion: ‘Coesite from the Richat Dome, Mauritania: A Misidentification.’ The mineral that had been reported as coesite was not coesite. With that identification withdrawn, the physical evidence for an impact went with it.

The same year, Fudali published alongside Robert Dietz and William Cassidy a second paper titled ‘Richat and Semsiyat domes (Mauritania): not astroblemes.’ Dietz had coined the term astrobleme for the eroded scar of an ancient impact, so this was a specialist ruling out a candidate in his own field. The paper’s reasoning is worth stating plainly because it is the kind of argument that can be checked: there was no structural evidence of the explosive shattering an impact produces, and no shock metamorphism in the minerals. The rocks had not been hit.

What these papers say is that the impact hypothesis failed on the physical evidence. What they do not say is that the structure is therefore mysterious, or that its origin was left open. They say a specific proposal was tested and did not survive the test. The coesite question itself did get a later chapter — Sharad Master and Joachim Karfunkel returned to it in a 2001 abstract in Meteoritics and Planetary Science, proposing an alternative origin for silica phases at the site — but by then the impact model had been dead for three decades.

What the Rings Are Made Of

The strongest argument against the Richat Structure being anything other than geology is not about craters at all. It is about stratigraphy, and it was settled in the early 1970s.

Michel Deynoux and Roland Trompette published a comparison in 1972 in the Bulletin de la Société Géologique de France, volume 13, pages 111 to 117, setting the rock sequence exposed inside the Richat against the sequence of the surrounding Adrar plateau. They match. The layers that form the concentric ridges are the same beds that run flat across the desert for hundreds of kilometres in every direction, bent upward into a dome here and then cut down by wind and water.

A year later, Théodore Monod and Charles Pomerol edited a volume collecting field studies of the site, published as Mémoire 28 in the series Sciences de la Terre under the title ‘Contributions à l’étude de l’accident circulaire des Richat.’ Monod was the outstanding desert naturalist of his generation, a man who crossed the Sahara on foot and by camel for decades and wrote down what he saw. The volume is a record of people standing on the rocks and describing them.

This is where a reader can test the artificial-construction idea themselves, without any specialist knowledge. If the rings were built, they would be made of construction material assembled by someone. They are instead made of the local bedrock, in the local order, with the same fossils and the same grain sizes as the plateau outside the structure. No excavation is required to check this. It is the finding of every published survey since the 1960s.

Dating the Uplift

The dome has been dated, more than once, by different methods and different teams.

Matton, Jébrak and Lee obtained an age of 98.2 million years, give or take 2.6 million, on hydrothermal breccia from the centre using argon-argon dating. Carbonatite from the structure had already been dated twice: Netto and colleagues reported 85 million years plus or minus 5 in 1992, and Poupeau and colleagues reported 99 million plus or minus 5 in 1996. The figures cluster in the Cretaceous.

Matton and Jébrak returned with a fuller model in 2014, in the Journal of African Earth Sciences, volume 97, pages 109 to 124, under the title ‘The eye of Africa (Richat dome, Mauritania): An isolated Cretaceous alkaline-hydrothermal complex.’ Their account has magma rising in stages — a bimodal tholeiitic series cut through by silica-poor carbonatitic and kimberlitic magmas — lifting the sedimentary cover into a dome, with a caldera collapse at the centre and hydrothermal fluids working through the fractured rock. Erosion then did the rest over tens of millions of years.

What the dating establishes is a formation history around a hundred million years old. What it does not establish, and what no paper on the site claims, is anything about the last twelve thousand years. That period is geologically uneventful here. Nothing in the record has this structure rising, sinking, or moving within any span of human history. Its Atlantic setting is not incidental to that story either, since the same continental break-up that opened the ocean is part of the tectonic context, a subject taken up in Plate Tectonics and the Geological Demise of a Sunken Atlantis.

Reading Critias Onto a Landform

The Atlantis identification is recent. It was set out in a 2018 YouTube series, picked up by British tabloids the same year, and has circulated steadily since. The case is almost entirely visual: Plato describes a city of alternating rings of land and water, and here is a ringed structure of roughly the right apparent scale.

Set that against what the dialogues actually contain. In the Timaeus, Atlantis is an island beyond the Pillars of Heracles, larger than Libya and Asia combined, and it sinks in a single day and night, leaving shoals that Plato says still obstructed shipping in his own time. In the Critias the capital is described in detail: a plain, concentric harbours dug by hand, bridges, a canal cut to the sea, walls faced in metal, and orichalcum.

The Richat Structure is four hundred metres above sea level and several hundred kilometres inland, on a stable stretch of African crust. It did not sink, and there are no shoals. Its rings dip outward, which is what erosion of a dome produces and not what a dug harbour looks like. There is no canal to the sea because there is no sea within reach. The scale is wrong in the other direction too, since a forty-kilometre feature is being offered as a city that Plato placed on an island bigger than two continents.

The habit of mapping the text onto whatever landscape resembles it is old, and it has produced a long line of candidate sites over five centuries, traced in Plato’s Shadow and the Age of Discovery: Early Modern Atlantis in Scholarship and Cartography. Where real historical events have been proposed as the seed of the story, they have generally been coastal catastrophes with physical traces, a line of argument set out in Echoes of Sunken Worlds: Unearthing Atlantis’s Real-World Inspirations.

The Stone Tools That Are Actually There

People did live at the Richat, and this is the part of the story most often left out.

A palaeoenvironmental study of the Quaternary sediments at Guelb er Richât by Ousmane Sao, Pierre Giresse, Henry de Lumley and colleagues examined the deposits in relation to Lower Palaeolithic sites in the surrounding Adrar. The archaeological material from the area is old: hand-axes and related stone tools of the kind associated with the Acheulean, made by people living in a Sahara that was periodically far wetter than it is now.

That is a genuine human record, and it is a far stranger one than a lost city. It means the rings were walked among by hominins whose toolkit predates writing, metal, agriculture and the wheel by an order of magnitude. It also means the site has been surveyed by people specifically looking for traces of occupation, which is why the absence of anything else carries weight.

What the archaeological record contains: stone tools, sediments, evidence of ancient lakes and wetter phases. What it does not contain: worked masonry, harbour works, metal, pottery of a state-level society, or any structure that required planning. The surveys were not searching for reasons to dismiss Atlantis. They were reconstructing climate and early human presence, and they reported what they found.

Everything in this article can be checked. The Fudali retraction is in Science, volume 166, issue 3902. The Matton papers are in Geology and the Journal of African Earth Sciences. Deynoux and Trompette are in the Bulletin de la Société Géologique de France. The Richat Structure does not need to have been a city to be worth the attention it gets. It is a hundred-million-year-old wound in the crust that a mineral misidentification once turned into a crater, and that patient fieldwork turned back into what it is.