Caral · Peru · -10.8933, -77.5206 · PYR record in the living atlas
The dates that moved American prehistory came out of the walls. When Ruth Shady, Jonathan Haas and Winifred Creamer published their radiocarbon results for Caral in Science in 2001, the material they dated was plant fibre from shicra: the woven reed carrying bags that Caral's builders filled with quarried stone and stacked inside their platform mounds. Fibre in a structural bag dates the act of building, not something dropped nearby afterwards. Six samples were processed, at Beta Analytic in Miami and at the Illinois State Geological Survey. The answer came back at around 2627 BC, with the monumental architecture running on the same calibration to roughly 1977 BC, and the history of urban America moved back by more than a thousand years in a single paper. What it displaced was the consensus that complex society in the Americas began with the Olmec.
The shicra bags were not only a convenience for hauling rubble. Stacked inside a platform, bags of loose stone give the mass a degree of flexibility under seismic load, on one of the most earthquake-prone coasts on Earth. The outer faces were finished in plastered stone. It is a construction method that reads as makeshift and behaves as engineering.
Whose city, and whose name
Shady had been working the Supe valley since 1994 and published her results in 1997, in Spanish, in Peru, where almost nobody outside the region read them. The 2001 paper in Science, with two American co-authors, went everywhere. Shady later accused Haas and Creamer of taking credit that was not theirs. Their institutions cleared them of plagiarism, and Haas was rebuked for giving too little credit to Shady in public material. The residue is still visible on any bookshelf: Spanish-language scholarship calls this civilisation Caral-Supe, English-language scholarship usually calls it Norte Chico. Two names for one set of ruins, and the split is about people rather than evidence. If you have only ever read about it in English, there is a fair chance you have met the second name and never the first.
Monuments before pots, and an argument still open
Caral is awkward for the standard sequence in which societies acquire pottery, then surplus, then monuments. It has monuments and no pottery at all. It sits roughly 20 to 25 kilometres inland from the Pacific. The Pirámide Mayor measures about 150 metres by 110 metres and stands somewhere between 28 and 30 metres high, depending on which account you read. The largest of the sunken circular plazas is about 100 metres across. The urban core covers something over 60 hectares, with published figures varying by a few hectares depending on where the boundary is drawn, inside a UNESCO-inscribed property of 626 hectares. There are at least eighteen large preceramic sites in the Supe valley, of which Caral is the largest; counts of nineteen and eighteen both circulate and the difference is only whether Caral itself is included in the total. Widen the frame to the four valleys of the Norte Chico and there are about 30 major centres inside roughly 1,800 square kilometres, which is a startling density for the first urban experiment on a continent.
Michael Moseley's maritime foundations hypothesis proposed that Andean civilisation was paid for by the sea, specifically by the enormous anchovy shoals of the Peruvian current, rather than by farming. Caral complicated that. It is inland, its people ate a great deal of cultivated plant food alongside their fish, and it is large. The version that has held up best is a trade in both directions: valley farmers grew cotton, coastal villages turned the cotton into nets and sent fish back up the valley, and the cities in the middle organised the exchange. Which end of that system came first, coast or valley, is genuinely unresolved, and the specialists say so.
Why so much of it is still there
Preservation at Caral is exceptional and there is a plain reason for it. The builders used no pottery, so surface collectors picking over the mounds found nothing worth carrying away, and the site escaped the heavy looting that has stripped so much of the Peruvian coast. UNESCO records only two later re-occupations, one around 1000 BC and one between roughly 900 and 1440 AD, neither of them systematic and neither of them built over the Preceramic architecture. Four thousand years later the platforms, the plazas and the bags inside the platforms are still where they were put. Excavation has continued without interruption under the Caral Archaeological Zone, a body of Peru's Ministry of Culture.
The absence everybody quotes, stated honestly
Caral has no fortifications, no battlements, no weapon caches, and no mutilated bodies. Neither does any other excavated Norte Chico site so far. That is the most quoted fact about this civilisation and it deserves to be quoted carefully, because it is an argument from absence across a region where most sites have never been dug. The honest form of the claim is not that warfare has been ruled out. It is that a century of work in four valleys has produced no convincing sign of it, over centuries of dense settlement, which is not what the same work produces almost anywhere else in the world. Excavation coverage across the wider network is incomplete, so the absence of weapons at the sites that have been dug cannot stand in for the whole of Norte Chico. That is a real result and a fragile one, and the next season of excavation could change it.
What Caral did leave behind in quantity is music. Excavators recovered a cache of flutes made from condor and pelican bone, usually reported as 32 though some accounts say 33, and 37 cornetts made from deer and llama bone. The flutes have been radiocarbon dated to 2170 BC plus or minus 90 years. There is also a knotted textile object identified as a quipu, the string recording device the Inca were still using four thousand years later. If the identification holds, it is by far the earliest example known. It is also a single object, thinly published, and it is the one Caral find we would treat with the most caution: extraordinary claims and single artefacts are a bad combination, and we would rather say that plainly than build a paragraph on it.
The Huanca, and what has actually been measured
Standing in front of one of the main pyramids is a monolith about 2.15 metres tall, known as the Huanca. The excavating team interpret it as a solar marker, used to track the time of day and the position of the sun through the year, and it is reported as sited due north of the structure it faces. Whether that placement is a deliberate and precise alignment or a by-product of the site grid has not been independently confirmed in peer-reviewed work.
What has been published is better than the usual offering at a site like this. In 2021 C. González-García, Ruth Shady Solís, J. Ricra, F. Criado-Boado and J. A. Belmonte published "The River and the Sky: Astronomy and Topography in Caral Society, America's First Urban Centers" in Latin American Antiquity. They examined the orientations of structures at Caral and at related Norte Chico sites against solar events and against features of the landscape, and found that both topography and celestial factors appear to have influenced where buildings were placed. That is a careful, measured, peer-reviewed finding with the site's own excavator as a co-author, and it is not the same claim as the popular one that Caral's layout encodes a calendar. Nobody has published that as a measured result.
Whether the monumental core was planned from the outset or grew across building phases is itself unresolved, including how the structures are oriented relative to each other, so a reader should hold the layout loosely.
One recent announcement should be recorded as an announcement. In 2026 the Caral Archaeological Zone reported a structure at the related coastal site of Áspero interpreted as an astronomical observatory. As of the middle of 2026 no peer-reviewed publication of the alignment measurements, the structural analysis or any statistical evaluation had appeared. We are noting it here precisely because it is preliminary, and because unverified announcements of this kind get repeated as findings within a year.
What Vichama drew on its own wall
Around 1800 BC the monumental centre at Caral was largely abandoned. Why is not a matter of speculation any more, because one of the Caral-Supe settlements left a picture of what was happening to it. At Vichama, on the coast, Shady's team uncovered adobe reliefs roughly 3,800 years old. One panel, about one metre high and 2.8 metres long, shows four human heads with their eyes closed, entwined by two serpents, and a fifth anthropomorphic head at the point where the serpents meet, this one wide-eyed, with five appendages spreading from it. Reliefs nearby show emaciated human figures. Shady reads the serpents as a water deity that waters the earth and makes seed germinate, and the central head as a sprouting seed. Whatever the theology, the subject matter is drought and hunger, drawn by the people living through it. Very few societies this old have left a record of their own crisis in their own imagery.
The leading explanation for the abandonment, argued by Shady on the basis of that work, is the prolonged drought associated with the climate event of roughly 4,200 years ago, with relocation rather than collapse as the outcome. That is now an evidenced hypothesis rather than a shrug. It is not a proven cause.
In July 2025 Peru opened Peñico to visitors after eight years of excavation and restoration led by Shady. It sits inland in the Huaura province, positioned where routes from the coast, the highlands and the Amazon converge, and it dates to roughly 1800 to 1500 BC, exactly the window in which Caral was emptying out. Eighteen structures have been excavated: ceremonial buildings, houses and a central plaza. The finds include clay figurines of people and animals, shell-bead necklaces, and pututus, trumpets cut from the shell of the Eastern Pacific giant conch. Material from the Amazon and from Ecuadorian waters turns up in the same deposits. The reading Shady's team offer is that the Caral-Supe world did not end at the 4.2 kiloyear climate event; it moved, rebuilt and kept its organisation, which is a different outcome from the disruptions the same episode caused in Mesopotamia and the Indus.
What oldest is actually claiming
Caral is routinely called the oldest city in the Americas and every word of that is load-bearing. Settlement in the Norte Chico valleys reaches back to around 3200 BC, several centuries before Caral's monumental phase, with the expansion peaking between 2500 and 2000 BC. Several of the 30-odd centres are contemporaries of Caral rather than descendants of it. Caral is the largest, the best preserved and by a wide margin the best excavated, and it is where the decisive dates were taken. Whether it was the first is a different question from whether it is the oldest one we can currently date, and nobody has answered the first one.
Two of the other words are contested as well. Whether "city" is the right term to carry across cultures at all is debated among the specialists, and so is whether Norte Chico is better read as a network of peer settlements than as a single polity with Caral at its head. Neither is a quibble about labels. They decide what the map of this civilisation is supposed to look like.
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. Caral returned 19 golden-ratio pairs where chance predicts 24.1, giving p = 0.6787.
That is fewer than chance would predict, and it is not significant. The control points the same way: the arbitrary ratio 2.0 scored 1.72 times its expected count against phi's 0.79, so nothing here singles out the golden ratio.
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.
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 Caral record in the living atlas to see the Supe valley in 3D, place the city among its neighbouring Norte Chico centres and run the geometry engines on the real coordinates. The Caral site-emblem tee carries the Mayor Pyramid in the Pyramason Chronicle style, with the site name and the exact coordinates.
Sources worth reading
Shady Solís, Haas and Creamer, "Dating Caral, a Preceramic Site in the Supe Valley on the Central Coast of Peru", Science (2001). Haas, Creamer and Ruiz, "Dating the Late Archaic occupation of the Norte Chico region in Peru", Nature (2004). González-García, Shady Solís, Ricra, Criado-Boado and Belmonte, "The River and the Sky: Astronomy and Topography in Caral Society, America's First Urban Centers", Latin American Antiquity (2021). How Caral communities survived a climate disaster, Archaeology Magazine, November 2025.