Yonaguni · Japan · 24.4360, 123.0110 · PYR record in the living atlas
In 1986 Kihachiro Aratake, who ran the Yonaguni tourism association and was in the water looking for hammerhead sharks, swam over something with corners. Off the southern tip of Japan's westernmost island, below the Arakawabana cliff, a mass of stone rises in sharp terraces and flat planes from the seabed to within a few metres of the surface. Photographs of divers hanging above its right-angled steps went around the world with a question attached: is this a monument drowned by rising seas at the end of the Ice Age?
How big is it, actually
Start with the measurements, because they are quoted more confidently than they deserve to be. The central terraced structure, the one in every photograph, is roughly 50 metres long and 20 metres wide, rising from about 26 metres depth to within about 5 metres of the surface. Figures of 300 by 150 metres also circulate widely and are often attached to "the main formation", which overstates its area by an order of magnitude. That larger figure describes the wider spread of features Masaaki Kimura mapped across the seabed, not the terraced block. And where the formation ends and ordinary seabed begins is itself part of what is disputed, which is why the two numbers can both be defended by the people quoting them.
Every dimension above comes from dive surveys rather than from precision survey. Nobody has run a metric survey of this thing to the standard that would let anyone argue about a metre.
The rock is 20 million years old and it fractures in squares
The formation is cut into the Yaeyama Group, a sequence of Early Miocene sandstones and mudstones roughly 20 million years old. Rock of that kind splits along two sets of surfaces at once: flat, near-horizontal bedding planes where one layer of ancient sediment sat on the next, and steep, straight joints opened by later tectonic movement. Where those two sets meet, weathering produces exactly what the photographs show, which is flat treads separated by clean vertical risers.
The strongest single argument against a monument is not an argument at all, it is a walk. At Sanninudai, on Yonaguni's own south coast, the same beds outcrop above water in the same staircase geometry, and nobody has ever called them ruins. That comparison is not casual. A 2019 study by Takayuki Ogata and colleagues, published in Island Arc, combined digital elevation models with field geology at three onshore geosites chosen for their resemblance to the submerged formation, Tindabana, Kube Ryofurishi and Sanninudai, and concluded that weathering and erosion acting on bedding planes and linear joints account for the shapes. Robert Schoch, a geologist better known for arguing that the Great Sphinx is older than Egyptology allows, dived the site and came away describing it as most likely natural: the walls as eroded platforms, the roads as joint-controlled channels, the drawings as scratches.
The other point the geologists make is structural rather than statistical. The formation is monolithic and continuous with the bedrock underneath it. Megalithic construction means blocks quarried somewhere and set somewhere else, and there is nothing on this formation that has been quarried, moved or placed.
Kimura's case, and the date he changed
The monument reading belongs to Masaaki Kimura, a marine geologist at the University of the Ryukyus, and it deserves stating at full strength rather than in caricature. Kimura has spent decades diving the site and describes a pyramid, a castle, roads, a stadium, drainage channels, post holes and a carved face. He reported the formation as megalithic construction to the Oceans '04 conference proceedings and again to the 21st Pacific Science Congress in 2007. He is not a man who read about it on the internet; he has been in the water more than anyone.
The other half of that assessment belongs here too, because leaving it out would be its own kind of dishonesty. Kimura's interpretation has been characterised as pseudoarchaeological in parts of the academic literature, and he has separately proposed the formation as a remnant of Mu, a lost Pacific continent for which there is no evidence of any kind. Both things are true at once: a credentialed researcher with more field time than his critics, arguing a position most of his field rejects, and attaching it to a claim that has no evidential basis. Neither half cancels the other.
The part almost every popular account omits is that Kimura moved his own date. His early estimate put the structure at 10,000 years or more, which is what tied it in the public mind to a drowned Ice Age civilisation. In his 2007 report he revised it to roughly 2,000 to 3,000 years ago, on the reasoning that sea level then was already close to today's and that tectonic subsidence, not glacial meltwater, put the rock under. That is the honest move of a researcher following his own evidence, and it also dissolves the framing that made Yonaguni famous. The Ice Age city is not Kimura's claim any more. It is still the headline everywhere else.
The substitution is worth looking at directly, though. The original date rested entirely on the site having stood above water before post-glacial sea-level rise, and that argument was the whole basis for it. Replacing the mechanism with tectonic subsidence does not repair the earlier reasoning so much as remove its foundation, and the two have never been squared with each other in print.
What has never been recovered
No excavated artefact, no datable cultural material and no documented tool mark has come off the formation. Kimura's reports of wedge marks have not been accompanied by published photographs of the marks on the structure with measured spacing, which is the ordinary standard for a claim of that kind. The face is visible from one vantage point and not from others, which is the signature of pareidolia rather than carving. Loose stone tools found nearby are as easily explained by centuries of fishing boats passing overhead.
That absence needs one caveat attached to it, and it usually is not. No artefact has been recovered in any published study, but nobody has systematically searched for one either. This is an underwater site in an active tectonic zone with strong currents, and no excavation of any kind has been carried out on it. Whether the absence of artefacts means nothing was ever there, or only that nobody has looked properly, is itself unknown.
The administrative fact worth knowing is that neither Japan's Agency for Cultural Affairs nor the Okinawa prefectural government treats the site as a cultural property, so no preservation work is done on it and no excavation licence governs it. That is also why this article has nothing new to report since 2019. Yonaguni has not moved in years, not because the question was answered but because nobody has dug and no official body intends to.
One popular claim can be dealt with quickly. The formation's main face is oriented roughly south-southeast, and no astronomical alignment for it has ever been measured or published by anyone. The honest form of that is not "there is no alignment" but "this has never been studied to a standard that would let any alignment claim be evaluated". Anyone quoting a bearing for Yonaguni is quoting nobody.
The third position
The debate is almost always reported as two camps, and there is a third that fits the evidence at least as well as either. Schoch, having concluded that the formation is natural, has also allowed that ancient people may have recognised a striking natural feature, used it as a sacred place, and possibly regularised steps that the rock had already produced. That would make it neither a building nor an untouched outcrop.
It is worth stating clearly because it is testable in principle and almost never mentioned. It would need artefact evidence, or tool marks documented to a publishable standard, or datable material in association with the formation, and none of those has been looked for. Until somebody looks, the position sits where it is: consistent with everything known, demonstrated by nothing.
Our own reading is that the geology accounts for the shapes without needing people, and that this is where most of the specialists who have examined the site come out. It is not a settled matter, it has not been tested by excavation, and the minority position is held by the person who has spent the most time on the site. If solid cultural evidence ever surfaces, we will update this article and be glad to. That is how the game is supposed to work.
The astonishing thing here is documented, and it is not the rock
People were in the southern Ryukyus a very long time ago, and the proof is on land. At the Shiraho Saonetabaru cave on Ishigaki, human remains about 27,000 years old have been recovered, the oldest known in Japan, along with the first Pleistocene burials identified in the country; the site was recommended for designation as a National Historic Site in November 2019. Getting there was the hard part, because the Kuroshio current runs between Taiwan and these islands like a river in the sea.
So a team led by Yousuke Kaifu of the University of Tokyo tested it. They felled a Japanese cedar and hollowed it with replicas of 30,000-year-old stone tools into a 7.5 metre dugout they named Sugime, then in 2019 paddled it 225 kilometres from eastern Taiwan to Yonaguni, more than 45 hours at sea. The results, with the current simulations that go with them, were published in Science Advances on 25 June 2025, and they show that the crossing works only if the paddlers aim well south-east of where they actually want to land. Yonaguni's real claim on our attention is that people reached it deliberately, in a hollowed log, across one of the strongest currents on the planet, tens of thousands of years ago. That story has dates, artefacts and a peer-reviewed test. The rock has none of the three.
What our engines found
Nothing usable, and the reason is ours rather than the site's. Our alignment engine measures great-circle distance ratios between catalogued sites. Inside a 200 km radius of Yonaguni our atlas holds only two catalogued sites. Two points give a single distance, far too few to tell signal from chance, so we report the gap instead of a number.
A sparse catalogue is a hole in our data. It is never evidence of absence in prehistory, and anyone who tells you otherwise is reading a limitation as a finding. Expanding coverage in this part of the world is on our roadmap, and when it lands this section will carry a real result you can reproduce yourself.
One limit worth stating plainly, because it is the one most often glossed over. This test compares positions between monuments. It says nothing about any building's own proportions, its height, base, slope or perimeter. That is a separate question, and at several sites a genuinely open one.
Explore it yourself
Open the Yonaguni record in the living atlas to see the formation's position off the island in 3D, place it among its neighbours and judge the coastline comparison yourself. The Yonaguni site-emblem tee carries the terraces in the Pyramason Chronicle style, with the site name and the exact coordinates.
Sources worth reading
Ogata, T. et al., "Topographical and geological analysis of Yonaguni Island geosites and the Yonaguni Monument", Island Arc, the peer-reviewed study behind the natural reading and the onshore comparison. Kimura's own reports to the Oceans '04 proceedings and to the 21st Pacific Science Congress (2007), which are where the monument case is actually made rather than repeated. Schoch's published account of his dives, for both the natural conclusion and the intermediate position. Hancock, Underworld (2002), listed here as the source of the popular claim and its reach, not as support for it.