StoneAtlas

Stone Atlas is a growing database of ancient megalithic sites, sourced and evidenced by you. Underneath it is the question nobody has settled: how were these built, and why.

Archaeologists, the people making videos about it, and whoever lives near a stone circle and has been photographing it for thirty years. The answers are already out there. Mostly they are buried in a comment section a week after somebody posts them.

Göbekli Tepe — T-Shaped ArchitectureTaula of Torralba — T-Shaped ArchitectureKarahan Tepe — T-Shaped ArchitectureNaveta des Tudons — T-Shaped ArchitectureOllantaytambo — Trapezoidal AperturesKarnak — Trapezoidal AperturesMachu Picchu — Trapezoidal AperturesTiwanaku — Trapezoidal AperturesNewgrange — Solstice & Equinox AlignmentAbu Simbel — Solstice & Equinox AlignmentChichén Itzá — Solstice & Equinox AlignmentStonehenge — Solstice & Equinox AlignmentAngkor Wat — Solstice & Equinox AlignmentStonehenge — Trilithon & Post-and-LintelĦaġar Qim — Trilithon & Post-and-LintelDolmen of Menga — Trilithon & Post-and-LintelMnajdra — Trilithon & Post-and-LintelBaalbek — Cyclopean WallsNan Madol — Cyclopean WallsCyclopean Walls of Argos — Cyclopean WallsAcropolis Cyclopean Walls — Cyclopean WallsHattusa — Cyclopean WallsUnfinished Obelisk — Precision Stone CuttingPuma Punku — Precision Stone CuttingYangshan Quarry — Precision Stone CuttingGreat Pyramid of Giza — Precision Stone CuttingNimrud — The Handbag SymbolLa Venta — The Handbag SymbolGöbekli Tepe — The Handbag SymbolCusco FoundationsOsaka Castle WallsSacsayhuamánPolygonal Wall of DelphiEdo Castle Walls

Polygonal Masonry WorldwideCusco FoundationsOsaka Castle Walls16,400 km · half a world apart

Walls built from irregularly shaped, many-sided blocks fitted tight enough to refuse a blade, laid without mortar and still standing through earthquakes that flattened later work beside them.

Independent answers to the same seismic problem, or a method that travelled? The catalog holds the examples; nobody has closed the question.

Tap any faded pin to jump to the Connection it belongs to.

ILLUSTRATION

Feature

Polygonal / Irregular Block Walls

Recorded at 5 sites

Language & sourcesOsaka Castle Walls

There is a Japanese survey of these wall techniques that has never been translated. I can summarise the typology it uses, chapter by chapter.

Someone who reads the local literature

The catalog

A Connection is a Feature plus the places it turns up.

Not a theory with sites attached to it. A shape someone can point at, a list of coordinates where it has been recorded, and an honest note about what that does and does not settle. Anyone can add the next coordinate.

ILLUSTRATION

Feature

Polygonal / Irregular Block Walls

16,400 km between the two furthest shown

Walls built from irregularly shaped, many-sided blocks fitted tight enough to refuse a blade, laid without mortar and still standing through earthquakes that flattened later work beside them.

Unsettled

Independent answers to the same seismic problem, or a method that travelled? The catalog holds the examples; nobody has closed the question.

Recorded at — 5 sites

  • Cusco Foundations

    Cusco Foundations

    Cusco, Peru

    13.52°S 71.98°W

  • Osaka Castle Walls

    Osaka Castle Walls

    Osaka, Japan

    34.69°N 135.53°E

  • Sacsayhuamán

    Sacsayhuamán

    Cusco, Peru

    13.51°S 71.98°W

  • Photo wanted

    Polygonal Wall of Delphi

    Phocis, Greece

    38.48°N 22.50°E

  • Edo Castle Walls

    Edo Castle Walls

    Tokyo, Japan

    35.68°N 139.75°E

Missing a site that belongs on this list? That is the contribution.

ILLUSTRATION

Feature

T-Shaped Forms

3,100 km between the two furthest shown

The monolithic T as a repeated architectural choice: the pillar tradition of Göbekli and Karahan Tepe, and the taula sanctuaries of Menorca — a shaped upright carrying a single broad capstone.

Unsettled

Eight millennia and the length of the Mediterranean separate them. That is the whole difficulty, and the reason both stay on the map together.

Recorded at — 4 sites

  • Göbekli Tepe

    Göbekli Tepe

    Şanlıurfa, Türkiye

    37.22°N 38.92°E

  • Photo wanted

    Taula of Torralba

    Menorca, Spain

    39.92°N 4.05°E

  • Karahan Tepe

    Karahan Tepe

    Şanlıurfa, Türkiye

    37.06°N 39.29°E

  • Naveta des Tudons

    Naveta des Tudons

    Menorca, Spain

    40.00°N 3.89°E

Missing a site that belongs on this list? That is the contribution.

ILLUSTRATION

Feature

Trapezoidal Apertures

12,100 km between the two furthest shown

Doorways, windows, and niches cut narrower at the top than at the base, giving openings a tapered profile that resists spreading.

Unsettled

A shape that any mason working in stone might arrive at — which is exactly why it needs counting rather than arguing about.

Recorded at — 4 sites

  • Ollantaytambo

    Ollantaytambo

    Sacred Valley, Peru

    13.26°S 72.26°W

  • Karnak

    Karnak

    Luxor, Egypt

    25.72°N 32.66°E

  • Machu Picchu

    Machu Picchu

    Cusco Region, Peru

    13.16°S 72.55°W

  • Tiwanaku

    Tiwanaku

    La Paz, Bolivia

    16.55°S 68.67°W

Missing a site that belongs on this list? That is the contribution.

ILLUSTRATION

Feature

Solstice / Equinox Alignment

16,000 km between the two furthest shown

Structures set so that light from the rising or setting sun on one particular day of the year enters a passage, strikes a chamber, or runs the length of an axis — Newgrange at midwinter, Abu Simbel's inner sanctuary twice a year, Chichén Itzá's stair shadow at equinox.

Unsettled

That the alignments are real is measurable and measured. What they were for, and how much precision each culture was actually reaching for, is where the argument starts.

Recorded at — 5 sites

  • Newgrange

    Newgrange

    Boyne Valley, Ireland

    53.69°N 6.48°W

  • Abu Simbel

    Abu Simbel

    Aswan, Egypt

    22.34°N 31.63°E

  • Photo wanted

    Chichén Itzá

    Yucatán, Mexico

    20.68°N 88.57°W

  • Stonehenge

    Stonehenge

    Wiltshire, England

    51.18°N 1.83°W

  • Angkor Wat

    Angkor Wat

    Siem Reap, Cambodia

    13.41°N 103.87°E

Missing a site that belongs on this list? That is the contribution.

ILLUSTRATION

Feature

Trilithon (Post and Lintel)

2,100 km between the two furthest shown

Two upright stones carrying a single horizontal lintel as one structural unit — framing a doorway, or standing free with nothing behind it. The simplest way to hold a gap open in stone, and the one that had to be invented before anything larger could be.

Unsettled

The most likely candidate on this list for pure convergence: it is what the material asks for. Which is exactly why it is worth counting rather than assuming.

Recorded at — 4 sites

  • Stonehenge

    Stonehenge

    Wiltshire, England

    51.18°N 1.83°W

  • Ħaġar Qim

    Ħaġar Qim

    Qrendi, Malta

    35.83°N 14.44°E

  • Dolmen of Menga

    Dolmen of Menga

    Antequera, Spain

    37.02°N 4.55°W

  • Photo wanted

    Mnajdra

    Qrendi, Malta

    35.83°N 14.44°E

Missing a site that belongs on this list? That is the contribution.

4 mILLUSTRATION

Feature

Cyclopean Walls

13,300 km between the two furthest shown

Walls built from blocks too large to be set by hand — multi-ton, roughly dressed, laid without mortar. The Greeks who inherited them assumed giants had done it, which is where the name comes from.

Unsettled

Moving the blocks is the whole question, and it is a question about labour, ramps and time rather than about mystery. The map's job is to record where it happened, and at what scale.

Recorded at — 5 sites

  • Baalbek

    Baalbek

    Beqaa Valley, Lebanon

    34.01°N 36.20°E

  • Nan Madol

    Nan Madol

    Pohnpei, Micronesia

    6.84°N 158.33°E

  • Photo wanted

    Cyclopean Walls of Argos

    Argolis, Greece

    37.63°N 22.73°E

  • Photo wanted

    Acropolis Cyclopean Walls

    Athens, Greece

    37.97°N 23.73°E

  • Hattusa

    Hattusa

    Çorum, Türkiye

    40.02°N 34.62°E

Missing a site that belongs on this list? That is the contribution.

ILLUSTRATION

Feature

Precision Stone Cutting

18,100 km between the two furthest shown

Flat faces, square interior corners and true edges cut in granite, diorite and andesite — the hardest stone available — with the tool marks still on the unfinished work. Quarries left mid-cut are the most informative sites of all, because they show the process rather than the result.

Unsettled

Experimental archaeology has reproduced a good deal of this with copper, sand and patience. How much, and how fast, is an open and testable question.

Recorded at — 4 sites

  • Unfinished Obelisk

    Unfinished Obelisk

    Aswan, Egypt

    24.08°N 32.90°E

  • Puma Punku

    Puma Punku

    Tiwanaku, Bolivia

    16.56°S 68.68°W

  • Yangshan Quarry

    Yangshan Quarry

    Nanjing, China

    32.07°N 119.00°E

  • Great Pyramid of Giza

    Great Pyramid of Giza

    Giza, Egypt

    29.98°N 31.13°E

Missing a site that belongs on this list? That is the contribution.

ILLUSTRATION

Feature

Handbag Symbol

12,500 km between the two furthest shown

A bag-like object with a rounded handle over a flat base, held at the side by standing figures — carved on Assyrian palace reliefs, on Olmec stelae, and among the pictograms on Göbekli Tepe's Pillar 43.

Unsettled

Convergent shorthand for a common object, or one motif carried between them? Recorded as a resemblance. Not filed as an answer.

Recorded at — 3 sites

  • Nimrud

    Nimrud

    Nineveh, Iraq

    36.10°N 43.33°E

  • La Venta

    La Venta

    Tabasco, Mexico

    18.10°N 94.04°W

  • Göbekli Tepe

    Göbekli Tepe

    Şanlıurfa, Türkiye

    37.22°N 38.92°E

Missing a site that belongs on this list? That is the contribution.

The open questions

Nobody has settled how most of this was built, or why.

So the questions stay questions. Each one carries a list of what it would take to settle it, and any site can be tested against that list. Some of these will turn out to be wrong. That is fine, and finding out which is most of the work.

Being on this list is not an endorsement. It means somebody stated a claim clearly enough that other people can go and check it.

Think one of these is nonsense? Say so on its page — an argument with a source attached is worth more here than agreement.

The part that only people can do

Your audience already writes these. They just have nowhere to land.

Under every video about an ancient site there is someone who lives twenty minutes away, someone whose grandfather worked the 1962 survey, someone who has cut stone for thirty years and can tell you how that joint was made. A week later the next upload buries all of it. Here it attaches to the place instead, and it is still there in five years.

Worked examples, not testimonials — nobody has said these yet. They are the shape of a contribution that earns its place on the record.

  • Local accessOllantaytambo

    I drive past this every week. Tell me which face you need and I will photograph it properly — scale bar, raking light, the lot.

    Someone who lives twenty minutes away

  • Family archiveNimrud

    My grandfather worked the 1962 survey here. There is a box of his slides in the attic showing sections that have since been re-pointed.

    Someone clearing out a grandparent's house

  • Trade expertiseSacsayhuamán

    That joint was not ground flat. You can read the scribe marks — the block was offered up, marked, taken down, cut back, and re-offered. I do it the same way.

    A stonemason of thirty years

  • CorrectionKarahan Tepe

    The date on this pin traces back to a 1998 popular article, not the excavation report. The report says something narrower. Here is the page.

    A reader who checked the citation

  • Field reportTiwanaku

    The lower course is not visible in any published photo — it is below the modern walkway. I can get you down there in the dry season.

    A guide who works the site

  • Language & sourcesOsaka Castle Walls

    There is a Japanese survey of these wall techniques that has never been translated. I can summarise the typology it uses, chapter by chapter.

    Someone who reads the local literature

  • Local accessTaula of Torralba

    Happy to walk the whole taula group with a tape measure and a compass this month if somebody tells me what to record.

    Someone on the island

  • Trade expertiseGöbekli Tepe

    Give me twenty overlapping photos with a scale in frame and I will turn this into a measurable model everyone here can argue over.

    A surveyor

  • MeasurementNewgrange

    I stood the whole sunrise with a compass and a timer. The light reaches the back recess for about seventeen minutes, not the four the guidebooks keep repeating.

    Someone who was there at dawn on the shortest day

  • CorrectionPuma Punku

    This is described as granite on the site page. It is andesite. That changes the hardness argument completely, and probably the quarry it came from.

    A geologist

  • ProvenanceBaalbek

    The image everyone shares of this is a 1970s postcard, reversed. The real face is the other way round, which is why the joint never matches the drawings.

    Someone chasing a photo backwards

  • Field reportNan Madol

    Half the walls are only reachable at low tide and most visitors never see them. I go out most weeks and can carry a camera for whoever needs it.

    A boat operator

  • Trade expertiseUnfinished Obelisk

    Those are not decorative grooves. That is where a channel was being cut and abandoned — you can see the direction of work and where they stopped.

    A quarry worker

  • Family archiveHattusa

    I have thirty years of negatives of this place, including the sections that were consolidated in the nineties. Somebody should scan them before I am gone.

    A retired site photographer

  • MeasurementStonehenge

    Ran the alignment against sunrise azimuth for 3000 BCE rather than today. It shifts by about a degree, and that is worth writing down on the pin itself.

    An amateur astronomer

  • Language & sourcesDolmen of Menga

    The Spanish report distinguishes two build phases here. Every English summary I can find collapses them into one, which is where the confusion starts.

    Someone who reads the excavation reports in the original

Not a mockup

The map is already running. Go and pull it around.

Coloured pins are the eight Connections above, with the lines between them. The grey ones are the rest of the catalog, and every one of them is named — hover to see which. Real records at real coordinates, which is also why every one of them can be argued with.

Where most people arrive from

The video starts it. This is where it carries on.

A video about a site gets a hundred thousand views and four thousand comments. Somewhere in there is the person who lives twenty minutes away, the one with their grandfather’s survey photographs, the one who cuts stone for a living and can tell you how that joint was made. None of it survives contact with the next upload.

Here those land on the site itself, next to the evidence and the open question they bear on, and they are still there in five years. The video is the front door. It is not the building.

It is not finished, and this page is not a pitch for a finished thing. It is what exists so far, shown honestly, ahead of opening it up.

If you make this kind of work

Your videos can anchor to the sites they cover, and the useful half of your comment section can land somewhere it survives. Nothing is gated, nothing about your presence here costs anything, and you can have any of it taken down by asking.