Were There Lost Cities in the Amazon?

Background

For most of the 20th century, mainstream archaeology treated the Amazon rainforest as a "counterfeit paradise": a landscape whose visible lushness masked soil too acidic and nutrient-poor to support anything beyond small, mobile bands of hunter-gatherers and shifting cultivators. On this view, any historical account describing large, dense, complex Amazonian societies had to be exaggeration, and the region's own scale and remoteness were treated as sufficient explanation for why no ruins had ever been found.

That assumption had a specific historical account to explain away. In 1541-42, Spanish conquistador Francisco de Orellana led the first European expedition to travel the length of the Amazon River, after separating from Gonzalo Pizarro's search for cinnamon and, by extension, the gold and cities the era's Spanish explorers were also chasing across South America, the same broad ambition behind the El Dorado legend. Orellana's expedition chronicler, the Dominican friar Gaspar de Carvajal, recorded stretches of riverbank, in places extending for many miles, where the party never lost sight of houses, and described roads, public squares, large communal buildings, and populations he compared to the cities of contemporary Europe. For more than four centuries, historians and archaeologists dismissed Carvajal's account as a friar's embellishment, inconsistent with the sparse populations later explorers, including Percy Fawcett in the 1920s, actually encountered.

The Terra Preta Evidence

The first serious crack in the "pristine wilderness" consensus came from soil science rather than ruins. Researchers studying patches of unusually dark, fertile soil scattered across the Amazon basin, known as terra preta ("dark earth"), found that it was not a natural formation but a deliberate human creation: pre-Columbian populations had enriched the region's naturally poor, acidic soil by working in charcoal, organic waste, and broken pottery over generations, in some cases beginning thousands of years ago. Terra preta remains dramatically more fertile than the surrounding soil today. Its existence directly undercut the core ecological argument against large Amazonian populations, that the land simply could not have fed them, without yet proving that cities on the scale Carvajal described had actually existed.

The 21st-Century Archaeological Evidence

Direct confirmation followed from remote-sensing archaeology. In 2003, anthropologist Michael Heckenberger published excavations at Kuhikugu, on Brazil's Upper Xingu River, documenting a network of pre-Columbian "garden cities": fortified towns organised around central plazas, connected by broad roads, and surrounded by managed orchards, that supported tens of thousands of people at their height, a finding examined in more detail on this site's Percy Fawcett page.

A larger-scale confirmation followed in 2022, when a team led by Heiko Prümers used airborne LIDAR, a laser-based mapping technique that can see through forest canopy to the ground beneath, to survey the Llanos de Mojos, a savannah-forest region in Bolivia's Amazon basin. The survey revealed 26 settlement sites built by the Casarabe culture between roughly AD 500 and 1400, including two major centres of 147 and 315 hectares, organised in a four-tiered hierarchy and linked by straight raised causeways running for several kilometres. The largest sites featured stepped platforms, conical pyramids up to 22 metres tall, and concentric defensive banks, controlling a combined settlement network across some 4,500 square kilometres, a scale and density of pre-Hispanic urbanism not previously documented anywhere in Amazonia.

Common Misconceptions

The Amazon findings are frequently conflated with the El Dorado legend, since both drew European explorers into the same broad region and both are now popularly described as "lost cities." They are different claims: El Dorado specifically concerns a gold-laden city or empire that Spanish explorers invented by inflating a real Muisca coronation ritual, and no such golden city has ever been found or is expected to be. The confirmed Amazon settlements are earthwork and timber urban networks, no gold-filled temples, that demonstrate genuine large-scale, complex society, a different and, on the evidence, better-supported kind of claim.

It is also often assumed that "no ruins were found" meant "no cities existed." The absence of stone architecture, unlike Mesoamerican or Andean sites such as Tenochtitlan or Cusco, reflects building material, not scale: Amazonian societies built primarily with earth, timber, and thatch, materials that decay and are reclaimed by forest within decades, which is precisely why LIDAR, capable of detecting subtle ground-surface earthworks invisible from the forest floor or from historical ground-level exploration, was necessary to find them at all.

Current Consensus

Archaeologists now broadly agree that parts of the pre-Columbian Amazon supported large, complex, densely populated societies, a reversal of the mid-20th-century consensus achieved through terra preta soil science and, more recently, LIDAR-based remote sensing at Kuhikugu and in the Casarabe territory. Estimates for the pre-contact Amazonian population, sometimes cited in the millions, remain more provisional and are still actively debated among specialists, since they depend on extrapolating from surveyed areas to the whole basin. What is well documented is the collapse: European-introduced diseases, particularly smallpox and measles, are estimated to have killed the large majority of the Amazon's indigenous population within a generation or two of sustained contact, which is the leading explanation for why the region appeared sparsely inhabited to Orellana's own later successors and to 19th- and 20th-century explorers like Fawcett.

Why This Mystery Endures

The Amazon lost-cities question endures because its resolution runs against the grain of most "lost city" stories on this site: rather than a legend being debunked, a dismissed eyewitness account was vindicated by evidence that took nearly five centuries to arrive. Carvajal's 1542 description sat unbelieved through most of the history of Amazonian archaeology, not because it was disproven but because the technology to check it, laser mapping capable of seeing through dense canopy, did not exist until the 21st century.

It also endures because the story is still being actively written. The Casarabe survey covered only a fraction of the Amazon basin's total area, and researchers expect further LIDAR surveys to reveal more settlement networks in the coming years, a genuinely open, ongoing line of discovery rather than a closed historical question. The claim sits alongside Atlantis and El Dorado as one of this site's core lost-city cases, and is the clearest example among the three of a legend resolving toward "true, in a more modest and different form than imagined" rather than toward pure myth. The Amazon lost-cities question is part of this site's broader lost cities coverage, itself part of the wider ancient civilisations cluster.