Lost of the Amazon The Amazon tropical forest is not as wild as it looks By Michael J. Heckenberger

hen established the in 1961, the re- W serve was far from modern civili- zation, nestled deep in the southern reaches of the vast Amazon forest. When I fi rst went to live KEY CONCEPTS with the , one of the reserve’s principal ■ To most people, the Amazon indigenous groups, in 1992, the park’s boundar- forest is the quintessential case ies were still largely hidden in thick forest, little of pure nature slowly being more than lines on a map. Today the park is sur- destroyed as humans intrude. rounded by a patchwork of farmland, its borders ■ In fact, what seems pristine has often marked by a wall of trees. For many out- itself been shaped by humans. siders, this towering green threshold is a portal, In some areas the forest is sec- like the massive gates of Jurassic Park, between ondary growth that took hold the present—the dynamic modern world of soy when native peoples were fi elds, irrigation systems and 18-wheelers—and wiped out by their encounters the past, a timeless world of primordial nature with Europeans. The author and society. and his colleagues have found Long before taking center stage in the world’s extensive pre-Columbian . environmental crisis as the giant green jewel of Communities had a self-similar or fractal structure in which global ecology, the Amazon held a special place houses, settlements and clus- in the Western imagination. Mere mention of its ters of settlements were orga- name conjures images of dripping, vegetation- nized in similar ways. choked jungles; cryptic, colorful and often dan- gerous wildlife; endlessly convoluted river net- ■ Thus, the history of the Amazon is rather more interesting than works; and Stone Age tribes. To Westerners, usually thought. The environ- Amazonian peoples are quintessential simple so- mental challenge is not only to cieties, small groups that merely make do with later the Amazon still grips the popular imagi- preserve unspoiled wilderness what nature provides. They have complex nation as nature at its purest, home to native but also to recover the tech- knowledge about the natural world but lack the peoples who, in the words of edi- niques of sustainable farming hallmarks of civilization: centralized govern- tor Sean Woods in October 2007, preserve “a and forestry that the ancestors ment, urban settlements and economic produc- way of life unchanged since the dawn of time.” of the region’s present inhabit- tion beyond subsistence. In 1690 John Locke fa- Looks can be deceiving. Hidden under the ants developed. mously proclaimed, “In the beginning all the forest canopy are the remnants of a complex pre-

—The Editors World was America.” More than three centuries Columbian society. Working with the Kuikuro, LUIGI MARINI

64 SCIENTIFIC AMERICAN October 2009 I have excavated a network of ancestral towns, past human intervention. By developing a mix of KUHIKUGU, known to archaeolo- villages and roads that once supported a popu- land uses, soil-enrichment techniques and long gists as site X11, is the largest lation perhaps 20 times its present size. Huge crop rotation cycles, the ancestors of the Kui- pre-Columbian town yet discov- swaths of forest have grown over the ancient set- kuro thrived in the Amazon despite its infertile ered in the Xingu region of the tlements, gardens, fi elds and orchards, which fell natural soils. Their accomplishments could in- Amazon. It housed 1,000 or more people and served as the into disuse when epidemics brought by - form efforts to reconcile the environmental and hub of a network of smaller an explorers and colonists decimated the native development goals of this region and other parts towns. For a key to this artist’s peoples. The region’s rich biodiversity refl ects of the Amazon. conception, see page 68.

www.ScientificAmerican.com SCIENTIFIC AMERICAN 65 “Nature Folk” scribed the earliest expeditions, became an in- The most famous person to go looking for lost stant classic in the fl edgling discipline of anthro- civilizations in the southern Amazon was Percy pology. The book set the tone for 20th-century Harrison Fawcett. The British adventurer studies of Amazonian peoples as small, isolated scoured what he called the “uncharted jungles” groups living in a delicate balance with the trop- for an ancient , in the Amazon, ical forest: “nature folk.” Later anthropologists replete with stone pyramids, cobbled streets and often viewed the forest environment as uni- alphabetic writing. His tales inspired Conan formly inimical to agriculture; the soil’s poor Doyle’s The and perhaps the Indi- fertility seemed to preclude large settlements or ana Jones movies. ’s gripping recent dense regional populations. By this reasoning, book, The of Z , retraced Fawcett’s path the Amazon of the past must have looked much before his disappearance in the Xingu in 1925. like the Amazon in recent times. Actually, fi ve German expeditions had al- But this view began to erode in the 1970s as ready visited the Xinguano people and lands. In scholars revisited early European accounts of the 1894 ’s book Unter den region, which talked not of small tribes but of

Naturvölkern Zentral Brasiliens, which de- dense populations. As Charles Mann’s best-sell- RETSECK GEORGE

[THE REGION] X17 Lake Tafununo

X6 Land of Legend Archaeological surveys and excavations have revealed that X18 parts of the Amazon were densely populated on the eve of European and colonization. The author’s team X22 focuses on the area inhabited by the Kuikuro people, located in the headwaters of the in the Brazilian state of . It was here that the famous British explorer Percy X11 Harrison Fawcett disappeared more than eight decades ago X35/36 X38 when looking for lost cities.

10 Kilometers Kuikuro study area Boundary of

Areas of complex societies Xingu Indigenous Park Pre-Columbian towns

Caracas

Paramaribo Cayenne Marajó Island Belém Bogotá Santarém Manaus Quito BRAZIL Xingu River

Salvador Upper Tapajós Llanos Mato de Moxos Grosso Brasília Lima

66 SCIENTIFIC AMERICAN October 2009 [WHAT THE REGION IS LIKE TODAY] WA Kuikuro woman fi shes with a basket in a pond adjacent to the The Kuikuro archaeological site X13 (Heulugihïtï). The Kuikuro people were active participants in the TFlutists and dancers commemorate the late Brazilian anthropologist Apoena author’s fi eldwork and co-authors on the resulting Meirelles during the Kuarup funerary journal articles. Theirs is one of the few Amazonian festival in August 2005. societies to maintain the full breadth of its culture— language, rituals, art—despite centuries of depopu- lation, and they continue to practice intensive agri- culture (mostly of manioc), fi sh farming, orchard production and forest management.

SThe Kuikuro preserve a way of life that predates 1492. The archaeological record SA Xinguano village, with a population of a couple of hundred people, shows that they organize their economy and settlements as their ancestors did. consists of large thatch houses around a central plaza.

ing book 1491 has eloquently described, the Archaeological research in several areas were heavily populated on the eve of along the , such as Marajó Island [THE AUTHOR] the European landings, and the Amazon was no at the mouth of the river and sites near the mod- exception. Gaspar de Carvajal, the missionary ern cities of Santarém and Manaus, has con- who chronicled the first Spanish expedition fi rmed these accounts. These societies interacted down the river, noted fortifi ed towns, broad, in far-flung systems of trade. Less is known

); ); well-kept roads and large numbers of people. about the southern peripheries of the Amazon, Carvajal wrote on June 25, 1542: but recent work in Llanos de Mojos in lowland dancing ( Bolivia and in the Brazilian state of Acre sug- We went among some islands which gests that they, too, supported complex societies. AP Photo AP we thought uninhabited, but after we got In 1720 Brazilian frontiersman António Pires de to be in among them, so numerous were Campos described a densely settled landscape in Michael J. Heckenberger has ) the settlements which came into sight … the headwaters of the Tapajós River, just west of done archaeology in the Xingu

); ERALDO); PERES region and elsewhere in the that we grieved . . . and, when they saw us, the Xingu:

aerial view aerial Brazilian Amazon since 1992, ( author there came out to meet us on the river most recently as a professor and over two hundred pirogues [canoes], that These people exist in such vast quan- at the University of . shing fi each one carries twenty or thirty Indians tity, that it is not possible to count their His work focuses on the social Reuters/Corbis and political organization woman woman and some forty ... they were colorfully settlements or villages, [and] many times and historical ecology of late decorated with various emblems, and in one day’s march one passes ten or prehistoric complex societies in they had with them many trumpets and twelve villages, and in each one of them the region, as well as changes ); RICKEY); ROGERS drums ... and on land a marvelous thing there are ten to thirty houses, and in these in indigenous societies follow- child ( to see were the squadron formations that houses there are some that are thirty to ing European colonization not just in Brazil but also in north- Corbis were in the villages, all playing instru- forty paces across ... even their roads eastern South America, the ments and dancing about, manifesting they make very straight and wide, and Caribbean and northeastern great joy upon seeing that we were pass- they keep them so clean that one fi nds not North America. At home, he GREGG NEWTON GREGG COURTESY OF MICHAEL HECKENBERGER J. ( ing beyond their villages. even a fallen leaf.... relaxes by cultivating bonsai.

www.ScientificAmerican.com SCIENTIFIC AMERICAN 67 [GEOGRAPHY] Pre-Columbian Towns

Surveys show that the Kuikuro’s ancestors reworked hundreds of square together in clusters, each of which functioned as a political unit. The kilometers of forest into productive agricultural land. Although individual settlements had a fractal organization; for example, clusters, towns and settlements were small by modern standards, they were packed close houses were all organized along the same roughly east-west axis.

GPS-mapped road Medium-size town Extrapolated road Small town/village Lake Managed Lamakuka Plaza Gardens forest Large walled town + Unknown size Ritual center Farmed area Canoes Dwelling Earthen curb

X17 X16 X26 Kuikuro Lake Tafununo Village X14 X15 Lake X6 Ipatse

Ipatse Stream Angahuku River X13 Ipatse X25 Cluster Orchard X40 Forest Log X18 Floodplain palisade Marsh X22 X19 Gardens Road to X35/36 X44 X20 X42 Kuhikugu Forest Cluster X45 Marsh X33 Lake Lamakuka X21 X23 X34

X38 X11 10 kilometers Artist’s conception X35/36 of Kuhikugu (X11) Road Orchard

An Ancient Walled Town by extensive ditches. The Villas Boas brothers— When I ventured to Brazil in the early 1990s to Brazilian indigenistas who were nominated for study the deep history of the Xingu, lost cities a Nobel Peace Prize for their part in creating the were the furthest things from my mind. I had Xingu park—had reported such earthworks near read Steinen but had barely heard of Fawcett. many villages. Although much of the vast Amazon basin was In January 1993, soon after I arrived in the archaeological terra incognito, it was unlikely Kuikuro village, the principal hereditary chief, that ethnographers, much less local Xinguanos, Afukaka, took me to one of the ditches at a site had missed a large monolithic center towering (X6) they call Nokugu, named for the jaguar over the tropical forests. spirit being thought to live there. We passed local Nevertheless, signs of something more elabo- men who were raising a huge fi sh weir across the rate than present-day settlements were all Angahuku River, which was already swelling

around. Robert Carneiro of the American Mu- from the seasonal rains. The ditch, which runs ) seum of Natural History in New York City, who over two kilometers, was two to three meters

lived with the Kuikuro in the 1950s, had sug- deep and more than 10 meters wide. Even though view aerial gested that their settled way of life and produc- I had expected to fi nd an archaeological land- tive agricultural and fi shing economy could sup- scape different from today’s, the scale of these

port communities 1,000 to 2,000 strong—sever- ancient communities and their constructions sur- LUIGI); MARINI ( al times the contemporary population of a few prised me. Kuikuro research assistants and I map hundred. He also cited evidence that indeed it spent the following months mapping it and other once had: a prehistoric site (designated “X11” in earthworks at the 45-hectare site.

our archaeological survey) that was surrounded Since that time, our team has studied numer- ( RETSECK GEORGE

68 SCIENTIFIC AMERICAN October 2009 ous other sites in the area, hacking more than 30 a grassy glade of towering palms marking the XINGU HISTORY kilometers of line-of-sight transects through the former plaza. I slowly spun and pointed along forest to map, examine and excavate the sites. the perfectly circular edge of the plaza, marked Radiocarbon dating indicates that people have lived in the Upper Many Kuikuro helped in one way or another, by a meter-high mound. The tall palms, I told Xingu for at least 1,500 years. and some became well versed in archaeology. him, had colonized the plaza centuries ago from At the end of 1993, Afukaka and I went back compost gardens in domestic areas. 6th century. The ancestors of to Nokugu so I could tell him what I had learned. Leaving the plaza to explore the surrounding today’s inhabitants moved in from We followed the contour of the site’s outer ditch neighborhoods, we came across large refuse the west. and stopped at an earthen bridge, where a major middens that closely resembled the one behind road we had uncovered passed over it. I pointed Afukaka’s own house. They were fi lled with bro- 13th century. Groups organized themselves into integrated clusters down the arrow-straight ancient dirt road, ken pots that he noted were exactly like those with a regional population estimated which was 10 to 20 meters wide and led to an- his wives used to process and cook manioc, at 30,000 to 50,000. other ancient site, Heulugihïtï (X13), about fi ve down to minute details. On a later visit, when we kilometers away. We crossed the bridge and en- were excavating a pre-Columbian house, the 1542. Spanish conquistador Francis- tered Nokugu. chief bent down in the central kitchen area, co de Orellana led the fi rst European The road, defined by low earthen curbs, popped out a big hunk of pottery, and corrobo- expedition down the Amazon, as chronicled by Gaspar de Carvajal. widened to 40 meters—the size of a modern rated my sense that the daily life of the ancient four-lane highway. After a couple of hundred society was much like today’s. “You’re right!” 18th century. Slave raids further meters, we passed over the inner ditch and Afukaka exclaimed. “Look here, a pot sup- decimated the Xinguano people. stopped to look at our recently fi nished excava- port”—an undagi, as the Kuikuro call it, used to tion trench, where we had found a funnel-shaped cook manioc. 1884. German anthropologist Karl footing for a tree-trunk palisade. Afukaka told These connections are what make the Xin- von den Steinen visited the Xingu and me a story of palisaded villages and raids in his guano sites so fascinating. They are among the estimated a population of 3,500. people’s distant past. few pre-Columbian settlements in the Amazon 1950s. Orlando, Cláudio and Leonar- As we moved farther into the ancient town, where archaeological evidence can be linked di- do Villas Boas led a campaign to we passed through patches of forest, scrub and rectly to present-day customs. Elsewhere, the in- found the Xingu reserve. The Xingua- open areas that now cover the site—the foot- digenous culture was completely wiped out, or no population was about 500.

COURTESY OF U.S. GEOLOGICAL SURVEY/GODDARD SPACE FLIGHT CENTER/NASA/LANDSAT 7 CENTER/NASA/LANDSAT FLIGHT SPACE SURVEY/GODDARD GEOLOGICAL U.S. OF COURTESY prints of diverse past activities. We emerged into the archaeological record is spotty. The ancient

[REMOTE SENSING] Ipatse Cluster Peeking under the Canopy X17

Kuhikugu Cluster Landsat satellite images, falsely colored to repre- sent different infrared bands, show a mottled X21 X6 texture that indicates the X34 forest has been shaped by humans, particularly in and around ancient settlements that the X13 author and his team have X11 Saturated identifi ed in ground X38 anthropogenic surveys. Forest in unmod- forest ifi ed areas tends to look more uniform. Studies of X18 soil properties and vege- X35 tation confi rm that the region was once exten- sively cultivated. X19

High X22 GPS-positioned modern path forest Roads: major minor

www.ScientificAmerican.com SCIENTIFIC AMERICAN 69 At the present walled town I showed Afukaka was much like mately east-west; in the chief’s house, his ham- his current village, with its central plaza and ra- mock is oriented in the same direction. When a rate, the south- dial roads, only it was 10 times larger. chief dies, he is also laid to rest in a hammock ern Amazon with his head to the west. From House to Polity This basic corporeal calculus applies on all forest will be “Palatial” is not the word that usually comes to scales, from houses to the entire Upper Xingu reduced to 20 mind to describe a pole-and-thatch house. Most basin. Ancient towns are distributed across the Westerners think “hut.” But the house that the region and interconnected by a lattice of precise- ) percent of its Kuikuro were building for the chief when I ly aligned roads. When I fi rst arrived in the area,

arrived in 1993 was massive: well over 1,000 it took weeks to map the ditches, plazas and view aerial original size ( square meters. It is hard to imagine that a house roads using standard archaeological techniques. over the next built like a giant, overturned basket without Beginning in 2002, we began using precise GPS,

decade. Indige- stone, mortar or nails could get any bigger. Even enabling us to map major earthworks in a mat- Reuters/Corbis the average Xinguano house, at 250 square ter of days. What we have found is an impressive nous people are meters, is as big as the average American home. degree of regional integration. The landscape What makes the chief’s house stand out is not planning seems almost overdetermined, with a the stewards of RICKEY); ROGERS

just size but also its position, located on the specifi c place for everything. Yet it was based on map the remaining southern point of the central circular plaza. As the same basic principles of the current village. biodiversity. one enters the village along the formal entry Main roads run east-west, secondary roads ra- road, high-ranking families live to the right diate out to the north and south, and smaller (south) and left (north). The arrangement repro- roads proliferate in other directions. duces, on a larger scale, the layout of an individ- We mapped two hierarchical clusters of towns ual house, whose highest-ranking occupant and villages in our study area [see map on page hangs his hammock to the right, along the long 68]. Each consists of a major ceremonial center

axis of the house. The entry road runs approxi- and several large satellite towns in precise orien- COURTESY OF MICHAEL HECKENBERGER J. (

An Island of Trees Indigenous lands have become the most important barrier to deforestation in many parts of the Amazon. The Xingu reserve was once deep in the tropical forest, but cleared landscapes of cattle ranches and soybean fi elds now impinge on it from all sides. For Xinguanos to provide for their growing population, do they, too, need to clear-cut the forest? The author’s work suggests no. Their ancestors’ model of land use, based on small towns and long agricultural rotation cycles, can support a substantially larger population. Forested areas Deforested areas (or not part of the Amazon forest) Deforested in 2003 Deforested in 2004

Xingu reserve

DEFORESTATION in Mato Grosso is among the fastest in Brazil—in 2004 fi ve hectares a minute.

70 SCIENTIFIC AMERICAN October 2009 tations relative to the center. These towns likely the true number could have been much higher. held 1,000 or more inhabitants. Smaller villages Radiocarbon dating of our excavated sites sug- are located farther from the center. The north- gests that ancestors of the Xinguanos settled the ern cluster is centered on X13, which is not a area, most likely from the west, and began to town so much as a ritual center, rather like a fair- mold the forests and wetlands to their design ground. Two large walled settlements lie equi- about 1,500 years ago or before. In the centuries distant to the north and south of X13, and two before Europeans fi rst discovered the Americas, medium-size walled towns lie equidistant to the the communities were re-formed into hierarchi- northeast and southwest. The southern cluster is cal clusters. Records date back only to 1884, so slightly different. It is centered on X11, which is the settlement patterns are our only way of esti- both a ritual center and a town, around which mating the pre-Columbian population; the scale are medium- and small-size plaza settlements. of the clusters suggests a regional population In land area, each cluster was more than 250 many times larger than today, perhaps number- square kilometers, of which about a fi fth was the ing 30,000 to 50,000. built-up core area, making it roughly equivalent in size to a small modern city. Today most of the Garden Cities of the Amazon ancient landscape is overgrown, but forests in A century ago Ebenezer Howard’s Garden Cit- the core areas have distinctive concentrations of ies of To-morrow proposed a model for low- certain plants, animals, soils and archaeological density, sustainable urban growth. A forerunner artifacts, such as prolifi c ceramics. Land use was of today’s green movement, Howard envisioned more intensive in the past, but the remains sug- networked towns as an alternative to an indus- gest that many practices were similar to those of trial world fi lling with high-rise cities. Ten towns the Kuikuro: manioc plots, small orchards of with tens of thousands of people, he suggested, pequi fruit trees and fi elds of sapé grass, the pre- could have the same functional and administra- ferred material for house thatch. The country- tive capacity of a single megacity. side was a patchy landscape interspersed with The ancient Xinguanos built such a system, a secondary forests that invaded fallow agricul- fl at, green style of urbanism or proto-urbanism: tural areas. Wetlands, which today are choked an inchoate garden city. Perhaps with Buriti palm, the most important industrial was in the right place but looking for the wrong crop, preserve diverse evidence of fi sh farming, thing: stone cities. What the small-scale centers such as artifi cial ponds, raised causeways and lacked in size and elaborate structures, they made ➥ weir footings. Outside the core areas was a more up for in numbers and integration. Had Howard MORE TO lightly populated green belt and even deeper for- known of them, he might have devoted a passage EXPLORE est wilderness between clusters. This forest, too, to the “Garden Cities of Yesterday.” The common Amazonia 1492: Pristine had its uses for animals, medicinal plants and conception of the city as a dense grid of masonry Forest or Cultural Parkland? certain trees, and it was considered the home of buildings dates to early desert oasis civilizations Michael J. Heckenberger et al. in diverse forest spirits. such as Mesopotamia but was uncharacteristic of Science, Vol. 301, pages 1710–1714; September 19, 2003. The areas in and around residential sites are many other environments. Not only the Ama- marked by dark earth, which the Kuikuro call zon’s tropical forests but also temperate forest The Ecology of Power: Culture, egepe, a highly fertile soil that has been enriched landscapes throughout much of medieval Europe Place and Personhood in the by household refuse and specialized soil-man- were dotted with towns and villages of similar Southern Amazon, AD 1000–2000. agement activities, such as controlled burning of size to those in the Xingu. Michael J. Heckenberger. Routledge, 2005. vegetation cover. People have altered soils the These insights are especially important today world over, making them darker, more loamy as the southern Amazon is redeveloped, this 1491: New Revelations of the and richer in certain chemicals. In the Amazon time by Western civilization. The transitional Americas before Columbus. these changes are particularly important for ag- forest of the southern Amazon is being quickly Charles C. Mann. Vintage, 2006. riculture in many areas because the natural soil converted into farmland and pasture. At the Pre-Columbian Urbanism, is so poor. In the Xingu, the dark earth is less present rate, it will be reduced to 20 percent of Anthropogenic Landscapes, and prevalent than some areas, because local popu- its original size over the next decade. Much of the Future of the Amazon. lations depend mostly on manioc and orchards, what is left will be restricted to reserves, such as Michael J. Heckenberger et al. in which do not require high-fertility soils. the Xingu, where indigenous people are the Science, Vol. 321, pages 1214–1217; Identifi cation of large walled settlements over stewards of the remaining biodiversity. In these August 29, 2008. an area about the size of Vermont suggests that areas, saving tropical forests and protecting in- The : A Tale of Deadly at least 15 clusters were spread across the Upper digenous cultural heritage are, in many respects, Obsession in the Amazon. Xingu. But most of the region is unstudied, so one and the same thing. ■ David Grann. Doubleday, 2009.

www.ScientificAmerican.com SCIENTIFIC AMERICAN 71