Lost Cities of the Amazon Found

September 21, 2003

Newly discovered traces of ancient roads, bridges, and plazas in Brazil’s tropical forest may help dispel the once-popular impression of an “untouched” Amazon before the Europeans’ arrival. In southern Brazil, archaeologists have found the remains of a network of urban communities that apparently hosted a population many thousands strong. Reporting their findings in the journal Science, published by AAAS, the science society, the researchers say the people who dwelled there dramatically changed their local landscape. IN THE UPPER Xingu region of the southern Amazon, in central Brazil, Michael Heckenberger of the University of Florida and his colleagues have discovered… (read more)

…centuries-old remains of roads that appear to link a network of large villages in a carefully organized, gridlike pattern.

The residents, ancestors of the modern-day Xinguanos, dug enormous ditches around the villages, built bridges and moats in wetland areas, and cultivated large tracts of land. It seems that virtually no part of this landscape was truly wild, or “pristine.” Even some of the forested areas may have been more akin to a large park than to untouched forest, according to Heckenberger.


Though multitudes of plants and animals thrive in the Amazon, the environment was long thought to be too hostile for large-scale human settlement. In particular, archaeologists believed that the soil quality was too poor to support the intensive agriculture that would be necessary to support a population of significant size.

The general impression of native Amazonians as “stone age primitives frozen at the dawn of time” has changed little over the past few centuries, Heckenberger said.

“There was this cherished image that the Amazon was pure nature. The problem is, we have very few good, empirical cases that tell us what Amazonia was like in 1492, one way or the other,” he said. In recent years, archaeologists have been revising their view of the Amazon, sometimes provoking bitter debates over how extensively the land could have been settled by humans. A key reason for the controversy has been the lack of good physical evidence, according to Heckenberger.

The first written record that refers to the Kuikuro, a subgroup of the Xinguanos with whom Heckenberger has worked for a decade, is from 1884. But, according to the Kuikuro’s oral history, the first Europeans they encountered were slavers, around 1750. Heckenberger and his colleagues tentatively estimate that the population of the region numbered in the tens of thousands, but crashed due to enslavement and disease epidemics. By the 1950s, there were as few as 500 Xinguanos.

With indigenous Amazonians’ numbers decimated, and little concrete evidence of their earlier civilizations, researchers visiting the Amazon generally concluded that its people had always been small, “primitive” tribes who left little imprint on their environment.

Heckenberger’s team has found 19 settlements to date, at least four of which were major residential centers. The settlements were built around large, circular plazas, with roads leading out from them at specific angles, repeated from one plaza to the next.

Heckenberger, who collaborated with two Kuikuro chiefs on the Science study, believes the engineered features of the landscape all involved elements of the Kuikuro’s understanding of the entire cosmos. Road directions and the orientations of other structures are keyed to the directions of the sun and stars, for example. Today, the Kuikuro continue this sort of “ethnocartography,” as Heckenberger calls it.

Roads in the ancient settlements were up to 165 feet (50 meters) wide, the width of a modern-day four-lane highway, and flanked by large curbs. The researchers report that the roads linked settlements, every two to three miles (three to five kilometers), along an extensive grid. This kind of planning would have required the relatively sophisticated ability to reproduce angles over large distances, according to Heckenberger.

Where the villages converged on wetlands, the researchers discovered the remains of ancient bridges, moats and canals. The Kuikuro still use many such structures today.

The entire area in between settlements was carefully engineered and managed, according to the researchers. It was likely either cultivated, or maintained as a sort of parkland — a managed area, rather than wild or pristine forest. Satellite images reveal that the vegetation now growing in these areas looks quite different from older forest.

The Upper Xingu is the largest contiguous tract of Amazonian forest still under indigenous management. Its history brings up the question of how to go about conserving the remaining Amazon. Should the goal be to preserve a “pristine” wilderness untouched by human activity? Or a working landscape that supports indigenous peoples?

Perhaps both options need not be mutually exclusive. Heckenberger is quick to point out that the Amazon is not a uniform landscape.

“Because it’s so poorly known, Western knowledge has tended to treat the area as one homogeneous thing: one big jungle, one big rainforest, one natural lab for primitive people,” Heckenberger said. “As we dig into the region, we realize that 500 years ago it was very different, and that even today there is a large amount of variation that we didn’t appreciate before.

“These people were involved in the same kinds of cultural human innovation as elsewhere in the world. We’re not talking about the Incan or Roman Empire here, but in terms of the rest of Europe, Asia, Africa, the Americas and elsewhere, Amazonians were no less capable of human cultural innovation than anyone else,” Heckenberger said.

By American Association for the Advancement of Science