Angkor Wat · Cambodia · 13.412487, 103.866683 · PYR record in the living atlas
In the first month of 1632 a Japanese traveller named Morimoto Ukondayu Kazufusa wrote on a wall of Angkor Wat that he had come to offer four Buddha images and to pray for the living and the dead. He believed he was standing in Jetavana, the monastery in India where the Buddha was said to have lived. His text is one of 14 Japanese ink inscriptions left in the temple between 1612 and 1632, 13 of them in the cruciform cloister and one on the wall of the north library. A Japanese plan of the building, studied by Yoshiaki Ishizawa and preserved in a copy made in 1715, still labels the place Jetavana. Two centuries before Europeans announced they had found Angkor Wat in the forest, pilgrims were signing its walls.
It was never lost
Angkor Wat covers 162.6 hectares inside its moat, which makes it the largest religious complex on Earth by land area. Published areas vary according to what is being counted: the moat's outer rectangle of roughly 1.5 by 1.3 kilometres works out nearer 195 hectares, and the figures around 200 hectares that circulate in the literature are measuring that rather than the ground inside the water. It was built in the first half of the twelfth century under Suryavarman II and dedicated to Vishnu, and its own name in the inscriptions was Vrah Vishnuloka, the sacred dwelling of Vishnu. Unlike almost every other monument in this atlas it was never abandoned. When the Khmer court left Angkor after the Ayutthayan attack of 1431 the temple changed faith rather than emptying out. Radiocarbon dating of the ponds inside the enclosure shows they went on being maintained; houses stood within the walls; Buddhist practice carried on without a break, and the seventeenth-century Japanese inscriptions record small settlements of foreign pilgrims living alongside Khmer residents. The image of explorers hacking a lost city out of the jungle flatters the explorers, not the facts.
Facing the wrong way
Almost every Angkorian temple faces east. Angkor Wat faces west, and that one fact carries an argument that has never been settled. West is the direction of the dead in Khmer tradition, which has long supported reading the temple as a funerary monument for Suryavarman II rather than, or as well as, the state temple of the empire. No royal cremation deposit has been confirmed inside the structure, so neither reading is closed off. This is a mainstream debate among Khmer specialists rather than an alternative-history claim, and it is worth knowing before anyone tells you flatly what Angkor Wat was built to do. What can be said without argument is that it was dedicated to Vishnu and that it faces the wrong way for a temple.
The stone did not come the long way round
Popular accounts move the sandstone roughly 90 kilometres: down a canal to the Tonle Sap lake, then back up the Siem Reap river to the site. Etsuo Uchida and Ichita Shimoda tested that on the ground and in satellite imagery and found more than 50 quarries at the south eastern foot of Mount Kulen, linked to Angkor by a much more direct waterway of about 34 kilometres. The blocks took the short route. It is a small correction with a large consequence, because a supply line a third the length changes the arithmetic of how fast the temple could go up, and Angkor Wat was substantially finished inside one reign. Work stopped when Suryavarman II died, which is why some of the bas-relief galleries were left unfinished.
The sandstone is only the visible fabric. Laterite, the local ironstone, was used extensively for foundations, inner walls and hidden structural fill, so a good deal of the building is not made of the stone people photograph. Three things about the construction are still open and are usually stated as though they were not: the size of the workforce, the duration of the campaign, and the method by which blocks were raised to height. The figures quoted for all three are extrapolations rather than records, and no Khmer source gives any of them.
A mountain made of water management
The five towers model Mount Meru, home of the gods, and the moat stands in for the cosmic ocean. It is 190 metres wide, more than 5 kilometres around, and encloses a rectangle 1.5 kilometres east to west by 1.3 kilometres north to south, with the temple's own outer wall measuring 1,024 by 802 metres inside it. The central tower is usually given as about 65 metres above the ground; that number is repeated everywhere and we have not been able to trace it back to a published survey, so treat it as the conventional figure rather than a measurement. The moat is also engineering. It holds the water table steady beneath a sand foundation so that the ground under the towers neither floods nor shrinks, which is a large part of why the building is still upright. Around it, airborne laser scanning has mapped the city Angkor Wat sat in. In 2021 Sarah Klassen and colleagues combined that lidar with excavation data, radiocarbon dates and a machine learning model of temple chronology, and published an estimate of 700,000 to 900,000 people living across the 3,000 square kilometre Greater Angkor Region at its twelfth to thirteenth century height.
Nearly every visible surface of the temple was carved. The galleries carry one of the longest continuous bas-relief programmes in the world: the Mahabharata and the Ramayana, the Churning of the Ocean of Milk, and historical scenes of armies on the march attributed to Suryavarman II. Sappho Marchal counted 1,737 female figures in 1926; Kent Davis's 2013 inventory, which includes worn, sealed-in and partly finished carvings, reached 1,796. The two counts differ by about sixty. Neither found two the same.
What the radar found under the lawn
In 2015 the journal Antiquity ran a group of papers on work inside the enclosure by Damian Evans, Roland Fletcher, Miriam Stark, Till Sonnemann and David Brotherson, and they changed the picture more than anything since the lidar itself. Ground-penetrating radar at the western entrance found the foundations of eight demolished towers, laterite and sandstone, cross-shaped with side porches, built at the very start of the project and pulled down deliberately once the outer wall and gate were complete. South of the moat the survey picked up a structure of rectangular sand spirals, reported in Antiquity as more than 1,500 by 600 metres and described elsewhere as a set of coils each about a kilometre long, raised in the mid to late twelfth century and used only briefly. Evans and Fletcher said plainly that how the spirals functioned is not at all clear, and that is still where it stands. It has no known parallel anywhere in the Angkorian world. Excavation of low mounds inside the walls turned up modest wooden houses with pottery and cooking ware, the homes of the people who serviced the temple rather than of officials.
Fletcher also identified unfinished earthworks that read as military: the only known case of an Angkorian temple being systematically modified for defence.
Miriam Stark's radiocarbon work indicates occupation on the site before the temple, by at least a century, though the character of that earlier phase is still poorly understood. It is worth being precise about that, because "there was something here before" is a sentence that gets stretched a long way in popular writing. What the dates point to is activity in the decades preceding a twelfth-century temple. Digging has not stopped either. In April 2020 archaeologists from the APSARA Authority working near the south western corner of the main structure recovered a cache of around 300 Buddha sculpture fragments spanning the twelfth to the nineteenth centuries, clustered in a way that suggests they were all put into the ground in a single event.
The drought at the end
Brendan Buckley's team built a 759-year climate record from Fokienia hodginsii trees in Bidoup Nui Ba National Park in Vietnam, about 700 kilometres from Angkor, covering 1250 to 2008. It contains two droughts either side of Angkor's decline: a long one through the 1330s to 1360s, and a shorter, sharper one in the 1400s to 1420s, with 1402 and 1403 the driest consecutive years in the whole record. Buckley's own framing is careful, and worth repeating rather than improving on: he would not say climate caused the collapse, but a 30-year drought had to have had an impact. Intense monsoons between and after the dry spells silted and scoured the canals, and a hydraulic city is only as good as its plumbing. Angkor Wat, which needs nothing but its own moat, came through it. The network around it did not.
The equinox claim, weighed
In 1976 Robert Stencel, Fred Gifford and Eleanor Moron argued in Science that Angkor Wat encoded a calendar, with a summer solstice sightline to the hill of Phnom Bok and a winter solstice line to the temple of Kuk Bangro. Giulio Magli went back over the case. Kuk Bangro, he points out, is not visible from Angkor Wat and in all probability never was, and its construction date is unknown, so it may well postdate the temple it is supposed to align with. The Phnom Bok geometry is real, but treating it as the reason the temple sits where it sits requires a motive that Khmer sources give no support for. What does survive the check is simpler and better: from the western gate at dawn on the equinoxes, the sun rises directly over the central tower. The orientation statistics are also unambiguous. Of 31 monuments measured, every one falls between 85 and 90 degrees, 19 of them between 89 and 90, and the deviations run only to the north of due east rather than scattering both ways. That is deliberate, and it does not need a significance test to say so. Magli's conclusion is that the temples were oriented on purpose and were not calendars in stone. We think that is the right place to leave it: an alignment you can stand and watch, and no good evidence for the numerology built on top of it.
One further body of work belongs here because it is peer-reviewed and because it gets cited as though it were settled. In 2019 William F. Romain published an analysis in the Journal of Skyscape Archaeology identifying more than seventy candidate solstice alignments across Angkor Wat and eleven nearby temples, arguing they were symbolic rather than observational. He separately proposed that Angkor Wat was sited where a summer solstice sunrise line to Rong Chen on Phnom Kulen crosses an east to west cardinal line to Preah Khan of Kompong Svay. The azimuths were derived from Google Earth Pro rather than from ground survey, a method that has drawn criticism within the field, and the results have not been independently corroborated by field measurement. Published and established are different words, and this is a good place to keep them apart.
What our engines found
We ran the same test here as everywhere else, with the settings fixed in advance and identical at every one of the fifty sites we cover: a 200 km search radius, at least 12 km between sampled sites, 15 sites, a 2 per cent tolerance, and 2,000 Monte Carlo runs against a null model resampled from real catalogued sites in the same region. The settings are frozen because the answer moves if you tune them, and tuning them after seeing the answer is exactly the trick we object to elsewhere. Angkor Wat returned 4 golden-ratio pairs where chance predicts 4.2, giving p = 0.5807 (run on 1 October 2026).
That is almost exactly what chance would predict, and it is not significant. The control points the same way: the neutral control ratio 1.35 scored 1.19 times its expected count against phi's 0.95, so nothing here singles out the golden ratio.
Worth flagging: the engine could reach only 8 catalogued sites here, so this result carries little weight in either direction. That is a limit on our data, not a verdict on the region.
We publish the misses because a tool that only reports its hits is not measuring anything. Every figure above is reproducible from the same coordinates, the same settings and the same seed. When the catalogue grows, the figures can move, so we rerun these tests and update this section rather than leave an old number standing.
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.
That limit bears directly on the section above. A test of distances between sites can neither support nor undermine a claim about which way a temple faces, or what rises over its central tower.
Explore it yourself
Open the living atlas to see Angkor among hundreds of catalogued Khmer temples, set the sky to a twelfth-century date and run the alignment and ley line engines on the real coordinates. The Angkor Wat Site-Emblem T-Shirt carries the five-tower silhouette, the site geometry, and the exact catalogued coordinates.
Sources worth reading
Evans and Fletcher, "The making of Angkor: fusing lidar and archaeological survey", PNAS (2013). Fletcher, Evans, Stark, Sonnemann and Brotherson in Antiquity (2015), for the ground-penetrating radar inside the enclosure. Stark in Antiquity (2019), for the occupation predating the temple. Klassen et al. (2021), for the population model. Romain, Journal of Skyscape Archaeology (2019), for the solstice case, read alongside Magli on the orientation statistics. Higham, The Civilization of Angkor (2001). Freeman and Jacques, Ancient Angkor (1999), still the best guide to what you are actually looking at on site. Uchida and Shimoda on the quarries and the short canal route.