Terracotta Army · China · 34.385, 109.273056 · PYR record in the living atlas
On 29 March 1974, Yang Zhifa, five of his brothers and their neighbour Wang Puzhi were digging an irrigation well in fields near Lintong, about 1.5 kilometres east of a large earth mound near Xi'an. They hit fired clay. They had broken into the eastern edge of a formation now estimated at more than 8,000 soldiers, 130 chariots with 520 horses and 150 cavalry horses, arranged in three pits. State archaeologists arrived quickly and systematic excavation began the same year. Pit 1 alone is 230 metres long and 62 metres wide, covering about 14,260 square metres, divided into eleven parallel corridors.
Most of the army is still in the ground, and it is worth being exact about how much. Across five decades of work, something over 2,000 warriors have actually been lifted, conserved and removed. Pit 1 by itself is estimated to still contain more than 6,000. Those two numbers get swapped constantly in secondary accounts, and the difference between them is most of the army. A third season of excavation in Pit 1 began in 2009 over roughly 430 square metres and had produced more than 220 figures, 16 horses, four chariots and a quantity of weapons by early 2023; work in Pit 2 resumed in 2015 and continues. Around 40,000 weapons have been recovered across the complex.
The figures in all three pits face broadly east, towards the territories of the states Qin conquered, and the usual reading is that they guard the dead emperor against threats from that direction. That is an observation about a bearing, not an astronomical claim, and the distinction matters here. No peer-reviewed study has demonstrated a precise measured astronomical alignment for any of the pits. Claims that the complex encodes stellar geography circulate anyway; we are naming them so that they can be refused, because without published bearings they are assertions rather than findings.
The faces, and what we actually know about them
The figures stand between roughly 1.75 and 2 metres tall, officers taller than infantry, and were assembled from workshop-made parts: legs, torsos, arms and heads produced separately and combined. The lower body was built up hollow from coils or slabs; heads and hands were modelled apart and attached. The heads came from a limited repertoire, at least ten face moulds, and were then finished by hand with applied clay for ears, hair, topknots and moustaches. That is how a workshop generates thousands of individual faces without modelling thousands of originals. Rank is legible in posture, armour detail and headdress rather than in the face.
Whether those faces portray particular men is a real question and it is not settled; it has been argued both ways for decades. In 2014 Andrew Bevan and colleagues published a study in the Journal of Archaeological Science that built three-dimensional models of the figures from ordinary photographs in order to bring morphometric comparison to bear on what they called a famous and contentious set of objects. A 2022 paper by Efremov and colleagues in npj Heritage Science took the same question to high-precision 3D point clouds and measured the facial features directly, which is at least a quantitative footing for an argument that has mostly been conducted by eye. The honest position is that the variation is real, its cause is unresolved, and anyone who tells you flatly that each warrior is the portrait of a named soldier has gone past the evidence. One small thing has never been explained: every warrior unearthed so far has single eyelids, and specialists are divided between reading that as ethnic convention and as a deliberate stylistic choice conveying martial severity.
They were also painted, and the paint is the most fragile thing on the site. A 2026 study in npj Heritage Science of officer figurines from pit K0006 found a two-layer system: a ground of urushi lacquer, then mineral pigments including cinnabar, malachite, azurite, a bone white and carbon black, over ceramic bodies of quartz-rich silty clay tempered with feldspar and fired between 975 and 1081 degrees Celsius. The lacquer ground is the problem. Exposed to Xi'an's dry air it can curl within about fifteen seconds and flake off within four minutes, which is why almost every warrior you have seen photographed is grey. The site is now engineering its way around that window: in September 2024 an emergency conservation laboratory was installed inside Pit 1 and a combined excavation-and-protection platform was activated at Pit 2, so that lifting, conserving and displaying can happen in the same place rather than at opposite ends of a journey.
The chromium myth
For decades the textbooks carried a marvel: the bronze weapons in the pits had stayed sharp because Qin metallurgists invented a chromium anti-rust treatment two thousand years before anyone else. In 2019 a team led by Marcos Martinón-Torres, with Li Xiuzhen, Xia Yin and colleagues, published a paper in Scientific Reports whose title states the result: surface chromium on these weapons is neither an ancient anti-rust treatment nor the reason for their good preservation. They analysed 464 weapon parts. Chromium appeared on 37 of them, about 8 per cent, and it clustered on the parts that had been in contact with lacquered wood, sword grips and crossbow stocks. The lacquer was rich in chromium. It migrated into the metal after burial.
The real preservative was ordinary. The burial soil is moderately alkaline, pH 8.1 to 8.5, very fine grained and low in organic content, and the bronze carries a high tin content, between 5 and 25 per cent, which develops a protective surface layer. We tell that story deliberately. A field that corrects its own showpiece finding in print is doing the thing we want more of, and the production system that replaced the myth is more interesting than the myth was.
The same test is worth applying to a newer story. Popular science writing has claimed that the Qin state deliberately engineered poisonous soil around the warrior pits to kill tomb robbers. The mercury anomaly around the burial mound is a real geophysical finding, discussed below. Attributing it to deliberate soil engineering around the pits, rather than to diffusion from the sealed central chamber, has no published archaeological support at all, and we refuse it on the same grounds we refused the chromium marvel: it sounds like ancient sophistication and there is no measurement behind it.
The men who marked the crossbows
In 2018 Andrew Bevan, Li Xiuzhen, Martinón-Torres and colleagues published a study in Heritage Science using multispectral photography on bronze crossbow triggers from Pit 1. Under the right wavelengths, painted marks appear that the eye cannot see: black characters reading jia, first of the ten Heavenly Stems, wu, martial, and jiu, nine. A lance scabbard carried a red mark reading sigong, the name of the Qin central workshop. The ink is soot based and probably made from pine.
They match across separate components of the same trigger, which points to parts made in the same or similar moulds by a particular workshop cell and assembled soon afterwards: small self-contained teams turning out complete units, rather than one long assembly line. The bone gives the same picture: a 2026 paper in Archaeometry examined 14 bone bow-end plates from Pit 1, identified the material by protein analysis as deer bone from its collagen peptides, and traced a standard sequence of file, chisel, abrasion and drilling marks under a lacquer coating. Tens of thousands of precision parts, made to a repeatable specification, by units that could be identified and held responsible.
Did the idea come from the west
The most persistent alternative reading of this site is that it is not entirely Chinese. Life-size naturalistic sculpture appears in China abruptly, with this emperor, and very little leads up to it. German Hafner suggested in 1986 that the tradition came from contact with the west. Lukas Nickel published the fullest version in the Bulletin of the School of Oriental and African Studies in 2013, and the archaeologist Duan Qingbo has pointed to resemblances between the figures and later Central Asian statuary at Khalchayan. Put at its strongest, the argument is not that Greeks built the army. It is that the idea of monumental figurative sculpture reached the Qin court along the corridors that later carried silk, and that a court capable of importing an idea did everything else itself.
The case against is substantial. The archaeologist Zhang Weixing has said there is no substantial evidence at all for the link. Li Xiuzhen, quoted in the press coverage that made the claim famous, said afterwards that she had been misrepresented, and that the warriors may be inspired by western culture but were uniquely made by the Chinese. Specialists at the Emperor Qinshihuang's Mausoleum Site Museum broadly dispute the transmission argument, noting that the figures postdate Alexander by only a generation and that the movement of a whole sculptural tradition across those distances has never been demonstrated. Johanna Hanink and Felipe Rojas have argued that this family of theories rests on old Eurocentric assumptions about which societies are capable of what. Our reading: the manufacturing evidence, the moulds, the workshop marks, the lacquer and the alloys, is thoroughly local, and the case for transmission currently rests on a resemblance rather than on an object, a name or a date. That is not nothing. It is also not enough. Unresolved rather than dismissed.
The tomb nobody has opened
Sima Qian wrote his account of the First Emperor roughly a century after the man died, and it is the nearest thing to a contemporary description that survives. It says the tomb held replicas of palaces and towers, rare objects, a hundred rivers made of mercury, representations of the heavenly bodies, and crossbows rigged to fire on intruders. It says 700,000 conscripted men worked on the mausoleum, a figure the writer John Man has pointed out is larger than the population of any city on Earth at the time. It is also unclear whether the number is meant to cover the warrior pits, the whole mausoleum precinct, or is the kind of round rhetorical figure ancient texts reach for, and no archaeology settles it. And Sima Qian never mentions the terracotta army at all, though he does record that the emperor cast monumental bronze statues for his palace. The largest sculptural project of Chinese antiquity is absent from the definitive history of the period, which may be the reason nobody thought to go looking for it for two thousand years.
The mound is still sealed. Its height is quoted two ways and both are correct about different things: it stands about 51.3 metres today, eroded down from an original commonly estimated at around 76 metres. The inner wall of the precinct runs 2.5 kilometres round and the outer wall 6.3 kilometres. Magnetic survey has traced a brick perimeter wall about 460 metres north to south by 390 metres east to west, and a main chamber roughly 80 by 50 metres and about 15 metres high, somewhere between 20 and 50 metres down. Soil probes have found mercury far above natural levels, with one sampling point reported at 1440 parts per billion against an average near 205 across 53 others, although some researchers attribute part of that signal to modern industrial pollution nearby rather than to the tomb. China has left the chamber closed until preservation technology improves, and the lacquer flaking off the warriors is the argument for waiting. The centre of the complex is unexcavated on purpose.
What is still open
The dating has no independent radiometric anchor. Nothing in the assemblage has been dated by a physical clock. The chronology rests on stratigraphic correlation with firmly dated Qin occupation layers, on the style of the objects, and on the texts. That is normal for a site of this period and it is not a weakness anyone is hiding, but it does mean the confident dates you see quoted are relative dates wearing absolute clothes.
The project was abandoned rather than finished. Work continued until roughly 208 to 209 BC, when the instability following the emperor's death interrupted it. A fourth pit, found during the early surveys, was empty. Why the work stopped where it did has not been established.
What happened to the pits soon afterwards is debated. Excavation in the tunnel beside Pit 1 has produced evidence suggesting that soldiers who had surrendered to the Chu warlord Xiang Yu attempted to loot or wreck the mausoleum. Whether that destruction was organised, and how far it went, is not agreed among specialists. The wooden roofs of the pits collapsed onto the figures below, which is why almost every warrior comes out of the ground in pieces.
And whether the whole conception had a predecessor, in Warring States burial practice, in mythologised accounts of earlier royal tombs, or in ideas arriving from outside, is discussed in Chinese scholarship without consensus.
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. Terracotta Army returned 1 golden-ratio pair where chance predicts 1.8, giving p = 0.8166 (run on 1 October 2026).
That is fewer than chance would predict, and it is not significant. The control points the same way: the neutral control ratio 1.35 scored 1.36 times its expected count against phi's 0.56, so nothing here singles out the golden ratio.
Worth flagging: the engine could reach only 6 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. It also says nothing about which way a pit faces, so the eastward orientation above is not something our engines can speak to.
Explore it yourself
See the Xi'an cluster in the living atlas, view the site record among its neighbours and run the alignment and ley line engines on the real coordinates. The Terracotta Army Site-Emblem T-Shirt carries a single warrior's face, the site geometry, and the exact coordinates of the buried army.
Sources worth reading
Martinón-Torres, Li Xiuzhen, Xia Yin et al., on the chromium result, Scientific Reports (2019), the paper that retired the marvel. Bevan et al., Journal of Archaeological Science (2014) on morphometrics of the faces, and Bevan, Li Xiuzhen, Martinón-Torres et al., Heritage Science (2018) on the crossbow workshop marks. Efremov et al., "Measurement and analysis of facial features of terracotta warriors based on high-precision 3D point clouds", npj Heritage Science (2022). Nickel, Bulletin of the School of Oriental and African Studies (2013) for the fullest version of the transmission argument. Portal, The First Emperor: China's Terracotta Army (British Museum Press, 2007). Shaanxi Institute of Archaeology, Excavation report on Pit 1 of the Qin terracotta warriors and horses, 1974 to 1984 (Cultural Relics Press, 1988), the primary excavation record.