Every fall & spring, & even in the winter when there is no snow, these large plains either in one place or other is constantly on fire, & when the Grass happens to be long & the wind high, the sight is grand & awful, & it drives along with amazing swiftness.

Fidler acknowledged that the fires could be "very dangerous" but understood their purpose. "These fires burning off the old grass," he observed, "in the ensuing Spring & Summer makes excellent fine sweet feed for the Horses & Buffalo, &c."

When Indian societies disintegrated from disease and mistreatment, forest invaded savanna in Wisconsin, Illinois, Kansas, Nebraska, Wyoming, and the Texas hill country. Europeans forgot what the landscape had looked like before and why. Captain John Palliser, traveling through the same lands as Fidler six decades later, lamented the Indians' "disastrous habit of setting the prairie on fire for the most trivial and worse than useless reasons." Afterward even the memory of indigenous fire faded. By the twentieth century biologists were stoutly denying its existence. The "open, park-like woods" seen by early settlers, Harvard naturalist Hugh Raup asserted in 1937, were not caused by fire; they "have been, from time immemorial, characteristic of vast areas in North America." Raup's summary description of the idea that they were due to regular, wide-scale Indian burning? "Inconceivable." "It is at least a fair assumption," a widely used college forestry textbook remarked in 1973, "that no habitual or systematic burning was carried out by Indians." In the western United States, the geographer Thomas R. Vale wrote in 2002, the "modest" Indian population "modified only a tiny fraction of the total landscape for their everyday living needs."

Vale is in the minority now. Spurred in part by historians like Cronon, most scientists have changed their minds about Indian fire. Using clever laboratory techniques, they have convinced themselves that in most cases the tribal lore and old chronicles were right all along: Indian embers were sparkling in the American night for centuries before the Sumerians climbed their ziggurats.

Carrying their flints and torches, Native Americans were living in balance with Nature-but they had their thumbs on the scale. Shaped for their comfort and convenience, the American landscape had come to fit their lives like comfortable clothing. It was a highly successful and stable system, if "stable" is the appropriate word for a regime that involves routinely enshrouding miles of countryside in smoke and ash. And it was a system that Indians were abandoning in ever-rising numbers at the time when Europeans came.

TEN THOUSAND MOUNDS.

Anyone who traveled up the Mississippi in 1100 A.D. would have seen it looming in the distance: a four-level earthen mound bigger than the Great Pyramid of Giza. Around it like echoes were as many as 120 smaller mounds, some topped by tall wooden palisades, which were in turn ringed by a network of irrigation and transportation canals; carefully located fields of maize; and hundreds of red-and-white-plastered wood homes with high-peaked, deeply thatched roofs like those on traditional Japanese farms. Located near the confluence of the Missouri, Illinois, and Mississippi Rivers, the Indian city of Cahokia was a busy port. Canoes flitted like hummingbirds across its waterfront: traders bringing copper and mother-of-pearl from faraway places; hunting parties bringing such rare treats as buffalo and elk; emissaries and soldiers in long vessels bristling with weaponry; workers ferrying wood from upstream for the ever hungry cookfires; the ubiquitous fishers with their nets and clubs. Covering five square miles and housing at least fifteen thousand people, Cahokia was the biggest concentration of people north of the Rio Grande until the eighteenth century.

Away from the riverside, Cahokia was hardly less busy and imposing. Its focal point was the great mound-Monks Mound, it is now called, named after a group of Trappists who lived nearby in the eighteenth and nineteenth centuries. Around its sides rushed a flow of men, their body paint and tattoos obscured by dust from the hardened, brick-like mud that lay underneath the entire city. Some built new mounds or maintained the old; others hauled wood for fuel and houses or carried water in leather pouches or weeded the maize fields with stone hoes. Women carried stacks of woven mats, baskets of fish and produce, yowling children. Cooksmoke chimneyed to the sky. Standards made of painted animal skins flapped everywhere. Anyone who has visited Siena or Venice knows how surprisingly noisy a city without engines can be. At peak times, given the right wind conditions, Cahokia must have been audible for miles.

Monks Mound opens onto a plaza a thousand feet long. In its southwest corner is a pair of mounds, one conical, one square. One day I climbed up their grassy sides at sunset. Hardly any other visitors were there. The humped outline of the vast heap of earth emerged from the empty green like a powerful prairie ship. The sun was low and the great mound was casting a shadow that looked long enough to reach the Allegheny Mountains. For a moment I saw no sign of contemporary life; St. Louis, just across the river, had not yet switched on its lights. Around me was the mound city and nothing but the mound city. To we moderns the sensation of being in a constructed environment is so ubiquitous as to be invisible-in the cocoon of our strip malls and automobiles, we are like the fish that cannot feel the water through which they swim. In Cahokia's day it was different. A thousand years ago it was the only place for a thousand miles in which one could be completely enveloped in an artificial landscape.

To visitors today it seems obvious that Cahokia and the many other mound sites in the Midwest and Southeast are the remains of Indian settlements. It did not seem so clear in the past. Nineteenth-century writers attributed the mound complexes to, among others, the Chinese, the Welsh, the Phoenicians, the lost nation of Atlantis, and various biblical personages. A widely touted theory assigned authorship to Scandinavian emigres, who later picked up stakes, moved to Mexico, and became the Toltecs. The science-fiction writer and archaeology buff Robert Silverberg devoted an entire entertaining book to the back-and-forth over the origin of the mounds, which intermittently preoccupied American intellectuals for a century. Thomas Jefferson removed a slice from a mound on his estate, examined the stratigraphic layers, and announced that Indians had made it. George Bancroft, one of the founders of American history, disagreed: the mounds, he wrote in 1840, were purely natural formations.

Charitably, one could say that Bancroft was correct: Cahokia was a product of its geography, which in turn was a product of the Ice Age. When the glaciers melted, water gushed south, creating the Mississippi River and the Illinois and Missouri Rivers that funnel into it. They met in a roil of water eighty miles wide. When the rivers receded, they exposed a wide strip of bottomland. Into this land a group of Indians coalesced sometime before 800 A.D.

Nobody knows what these people called themselves or which language they spoke. They were not "Cahokians"-that name, itself a linguistic garble, comes from an unrelated group that migrated to the area almost a thousand years later. Archaeologists are unlikely to find a better name, though. According to William Woods, the geographer and archaeologist at the University of Kansas, Monks Mound completely covers whatever habitation these people had before they built Cahokia. To see the remaining traces of this early settlement, scientists would have to jack up the whole enormous pile and dig underneath. Almost all that can be known with certainty about this initial group is that it belonged to a diverse, four-thousand-year-old tradition characterized by the construction of large earthen mounds.

Based around the Mississippi and its associated rivers, these societies scattered tens of thousands of mounds from southern Canada and the Great Plains to the Atlantic coast and the Gulf of Mexico. They were especially concentrated in the Ohio Valley, but nearly as many are found in the Southeast. Highways, farms, and housing developments have destroyed most of them, and scientists have investigated only a small fraction of the survivors. Most of the earthworks were shaped like big cones and stepped pyramids, but some were sculpted into enormous birds, lizards, bears, long-tailed "alligators," and, in Peebles, Ohio, a 1,330-foot-long serpent.

The earliest known examples appeared in northeastern Lousiana about 5,400 years ago, well before the advent of agriculture. For reasons unknown, Indians heaved up a ring of eleven irregularly sized mounds, most of them connected by a ridge, on a hill overlooking the course of the Ouachita River. The biggest was as tall as a two-story house. About a dozen similar sites are known, of which the Ouachita ring is the oldest and biggest. None of the mounds in any of these places cover burials or contain artifacts or show signs of use. Indeed, they seem to have so little purpose that archaeologist Joe Saunders of Northeast Louisiana University, whose team excavated the Ouachita mounds in 1997, half-jokingly speculated to Science that the motive for building them could have been the act of construction itself. "I know it sounds awfully Zen-like," he conceded.

The Ouachita mounds as they may have appeared at their creation, 5,400 years ago.

Because modern-day hunter-gatherers in Africa live in egalitarian bands that constantly move from place to place, archaeologists assumed that Native American hunter-gatherers must also have done so. Discovering the Louisiana mounds upset this view: they suggest that at least some early Indians were stay-at-homes. More important, they testify to levels of public authority and civic organization rarely associated with nomads. Building a ring of mounds with baskets or deerskins full of dirt is a long-term enterprise. During construction the workers must eat, which in turn means that other people must provide their food. Such levels of planning are ordinarily thought to kick in with the transition to agriculture. When people till and sow the land, anthropologists say, they set up systems to protect their investment. Eventually somebody ends up in charge of allocating goods and services. But the mound builders in Louisiana built these massive constructions at a time when agriculture was barely under way-it was like the whiff in the air from a faraway ocean. In the central river valleys of North America, people had a way of life without known analogue.

After these first mounds the record is sparse. After the Ouachita mounds, there is a gap in the record of more than a millennium. The curtain parts again in about 1500 B.C., when an archipelago of villages, the largest known as Poverty Point, grew up in the northeast corner of Louisiana. Located fifty-five miles from the Ouachita site, Poverty Point had as a focus a structure resembling an amphitheater: six concentric, C-shaped ridges, each five feet tall, on a bluff facing the river. The jaws of the widest C are 3,950 feet apart, an expanse so big that scientists did not recognize the ridges as constructions until they took aerial photographs of the site in the 1950s.

Now another gap: seven hundred years. The next major sequence occurs mainly in the Ohio Valley, hundreds of miles north. Here was a group known as the Adena-the name is that of a well-known site. Because Adena mounds served as tombs, researchers know more about their deaths than their lives. Accompanying the noble few in the tombs to the world of the deceased were copper beads and bracelets, stone tablets and collars, textiles and awls, and, sometimes, stone pipes in the shape of surreal animals. The head of the creature faced the user, who sucked in tobacco smoke from its mouth. It is widely believed that Adena tobacco was much stronger than today's tobacco-it was psychoactive.

Tobacco was only one of the crops grown at Adena villages. The Mississippi and Ohio Valleys and much of the U.S. Southeast were home to what is known as the Eastern Agricultural Complex. A full-fledged agricultural revolution with a multifarious suite of crops, the complex is an example of a major cultural innovation that has completely disappeared. Its crops were such unfamiliar plants as marshelder, knotweed, maygrass, and little barley. All of these species still exist; one could stock a specialty restaurant with them. (Sample menu: maygrass patties, steamed knotweed beans, and buffalo tongue.) No one seems to be doing that, though. In fact, farmers today treat several of these crops as weeds-they routinely blast little barley with herbicides. Archaeologists have tentative indications of early domestication in spots from Illinois to Alabama by 1000 B.C. But agriculture did not begin to flower, so to speak, until the Adena.

Adena influence in customs and artifacts can be spotted in archaeological sites from Indiana to Kentucky and all the way north to Vermont and even New Brunswick. For a long time researchers believed this indicated that the Adena had conquered other groups throughout this area, but many now believe that the influence was cultural: like European teenagers donning baggy pants and listening to hip-hop, Adena's neighbors adopted its customs. Archaeologists sometimes call the area in which such cultural flows occur an "interaction sphere." Both less and more than a nation, an interaction sphere is a region in which one society disseminates its symbols, values, and inventions to others; an example is medieval Europe, much of which fell under the sway of Gothic aesthetics and ideas. The Adena interaction sphere lasted from about 800 B.C. to about 100 B.C.

MOUNDBUILDERS, 3400 B.C.1400 A.D.

Textbooks sometimes say that the Adena were succeeded by the Hopewell, but the relation is unclear; the Hopewell may simply have been a later stage of the same culture. The Hopewell, too, built mounds, and like the Adena seem to have spoken an Algonquian language. ("Hopewell" refers to the farmer on whose property an early site was discovered.) Based in southern Ohio, the Hopewell interaction sphere lasted until about 400 A.D. and extended across two-thirds of what is now the United States. Into the Midwest came seashells from the Gulf of Mexico, silver from Ontario, fossil shark's teeth from Chesapeake Bay, and obsidian from Yellowstone. In return the Hopewell exported ideas: the bow and arrow, monumental earthworks, fired pottery (Adena pots were not put into kilns), and, probably most important, the Hopewell religion.

The Hopewell apparently sought spiritual ecstasy by putting themselves into trances, perhaps aided by tobacco. In this enraptured state, the soul journeys to other worlds. As is usually the case, people with special abilities emerged to assist travelers through the portal to the numinous. Over time these shamans became gatekeepers, controlling access to the supernatural realm. They passed on their control and privileges to their children, creating a hereditary priesthood: counselors to kings, if not kings themselves. They acquired healing lore, mastered and invented ceremonies, learned the numerous divinities in the Hopewell pantheon. We know little of these gods today, because few of their images have endured to the present. Presumably shamans recounted their stories to attentive crowds; almost certainly, they explained when and where the gods wanted to build mounds.

In the context of the village, the mound, visible everywhere, was as much a beacon as a medieval cathedral. As with Gothic churches, which had plazas for the outdoor performance of sacred mystery plays, the mounds had greens before them: ritual spaces for public use. Details of the performances are lost, but there is every indication that they were exuberant affairs. "There is a stunning vigor about the Ohio Hopewell...," Silverberg wrote, a flamboyance and fondness for excess that manifests itself not only in the intricate geometrical enclosures and the massive mounds, but in these gaudy displays of conspicuous consumption [in the tombs]. To envelop a corpse from head to feet in pearls, to weigh it down in many pounds of copper, to surround it with masterpieces of sculpture and pottery, and then to bury everything under tons of earth-this betokens a kind of cultural energy that numbs and awes those who follow after.

Vibrant and elaborate, perhaps a little vulgar in its passion for display, Hopewell religion spread through most of the eastern United States in the first four centuries A.D. As with the expansion of Christianity, the new converts are unlikely to have understood the religion in the same way as its founders. Nonetheless, its impact was profound. In a mutated form, it may well have given impetus to the rise of Cahokia.

THE RISE AND FALL OF THE AMERICAN BOTTOM.

Cahokia was one big piece in the mosaic of chiefdoms that covered the lower half of the Mississippi and the Southeast at the end of the first millennium A.D. Known collectively as "Mississippian" cultures, these societies arose several centuries after the decline of the Hopewell culture, and probably were its distant descendants. At any one time a few larger polities dominated the dozens or scores of small chiefdoms. Cahokia, biggest of all, was preeminent from about 950 to about 1250 A.D. It was an anomaly: the greatest city north of the Rio Grande, it was also the only city north of the Rio Grande. Five times or more bigger than any other Mississippian chiefdom, Cahokia's population of at least fifteen thousand made it comparable in size to London, but on a landmass without Paris, Cordoba, or Rome.

I call Cahokia a city so as to have a stick to beat it with, but it was not a city in any modern sense. A city provides goods and services for its surrounding area, exchanging food from the countryside for the products of its sophisticated craftspeople. By definition, its inhabitants are urban-they aren't farmers. Cahokia, however, was a huge collection of farmers packed cheek by jowl. It had few specialized craftworkers and no middle-class merchants. On reflection, Cahokia's dissimilarity to other cities is not surprising; having never seen a city, its citizens had to invent every aspect of urban life for themselves.

Despite the nineteenth-century fascination with the mounds, archaeologists did not begin to examine Cahokia thoroughly until the 1960s. Since then studies have gushed from the presses. By and large, they have only confirmed Cahokia's status as a statistical outlier. Cahokia sat on the eastern side of the American Bottom. Most of the area has clayey soil that is hard to till and prone to floods. Cahokia was located next to the largest stretch of good farmland in the entire American Bottom. At its far edge, a forest of oak and hickory topped a line of bluffs. The area was little settled until as late as 600 A.D., when people trickled in and formed small villages, groups of a few hundred who planted gardens and boated up and down the Mississippi to other villages. As the millennium approached, the American Bottom had a resident population of several thousand. Then, without much apparent warning, there was, according to the archaeologist Timothy R. Pauketat of the University of Illinois at Urbana-Champaign, what has been called a "Big Bang"-a few decades of tumultuous change.

Cahokia's mounds emerged from the Big Bang, along with the East St. Louis mound complex a mile away (the second biggest, after Cahokia, though now mostly destroyed) and the St. Louis mounds just across the Mississippi (the fourth biggest). Monks Mound was the first and most grandiose of the construction projects. Its core is a slab of clay about 900 feet long, 650 feet wide, and more than 20 feet tall. From an engineering standpoint, clay should never be selected as the bearing material for a big earthen monument. Clay readily absorbs water, expanding as it does. The American Bottom clay, known as smectite clay, is especially prone to swelling: its volume can increase by a factor of eight. Drying, it shrinks back to its original dimensions. Over time the heaving will destroy whatever is built on top of it. The Cahokians' solution to this problem was discovered mainly by Woods, the University of Kansas archaeologist and geographer, who has spent two decades excavating Monks Mound.

To minimize instability, he told me, the Cahokians kept the slab at a constant moisture level: wet but not too wet. Moistening the clay was easy-capillary action will draw up water from the floodplain, which has a high water table. The trick is to stop evaporation from drying out the top. In an impressive display of engineering savvy, the Cahokians encapsulated the slab, sealing it off from the air by wrapping it in thin, alternating layers of sand and clay. The sand acts as a shield for the slab. Water rises through the clay to meet it, but cannot proceed further because the sand is too loose for further capillary action. Nor can the water evaporate; the clay layers atop the sand press down and prevent air from coming in. In addition, the sand lets rainfall drain away from the mound, preventing it from swelling too much. The final result covered almost fifteen acres and was the largest earthen structure in the Western Hemisphere; though built out of unsuitable material in a floodplain, it has stood for a thousand years.

Reconstructions of Cahokia, the greatest city north of the Rio Grande, ca. 1250 A.D.

Because the slab had to stay moist, it must have been built and covered quickly, a task requiring a big workforce. Evidence suggests that people moved from miles around to the American Bottom to be part of the project. If the ideas of Pauketat, the University of Illinois archaeologist, are correct, the immigrants probably came to regret their decision to move. To his way of thinking, the Big Bang occurred after a single ambitious person seized power, perhaps in a coup. Although his reign may have begun idealistically, Cahokia quickly became an autocracy; in an Ozymandiac extension of his ego, the supreme leader set in motion the construction projects. Loyalists forced immigrants to join the labor squads, maintaining control with the occasional massacre. Burials show the growing power of the elite: in a small mound half a mile south of Monks Mound, archaeologists in the late 1960s uncovered six high-status people interred with shell beads, copper ornaments, mica artworks-and the sacrificed bodies of more than a hundred retainers. Among them were fifty young women who had been buried alive.

Woods disagrees with what he calls the "proto-Stalinist work camp" scenario. Nobody was forced to erect Monks Mound, he says. Despite the intermittent displays of coercion, he says, Cahokians put it up "because they wanted to." They "were proud to be part of these symbols of community identity." Monks Mound and its fellows were, in part, a shout-out to the world-Look at us! We're doing something different! It was also the construction of a landscape of sacred power, built in an atmosphere of ecstatic religious celebration. The American Bottom, in this scenario, was the site of one of the world's most spectacular tent revivals. Equally important, Woods says, the mound city was in large part an outgrowth of the community's previous adoption of maize.

Before Cahokia's rise, people were slowly hunting the local deer and bison populations to extinction. The crops in the Eastern Agricultural Complex could not readily make up the difference. Among other problems, most had small seeds-imagine trying to feed a family on sesame seeds, and you have some idea of what it would have been like to subsist on maygrass. Maize had been available since the first century B.C. (It would have arrived sooner, but Indians had to breed landraces that could tolerate the cooler weather, shorter growing seasons, and longer summer days of the north.) The Hopewell, however, almost ignored it. Somewhere around 800 A.D. their hungry successors took another look at the crop and liked what they saw. The American Bottom, with its plenitude of easily cleared, maize-suitable land, was one of the best places to grow it for a considerable distance. The newcomers needed to store their harvests for the winter, a task most efficiently accomplished with a communal granary. The granary needed to be supervised-an invitation to develop centralized power. Growth happened fast and may well have been hurried along by a charismatic leader, Woods said, but something like Cahokia probably would have happened anyway.

THE AMERICAN BOTTOM, 1300 A.D.

Maize also played a role in the city's disintegration. Cahokia represented the first time Indians north of the Rio Grande had tried to feed and shelter fifteen thousand people in one place, and they made beginner's mistakes. To obtain fuel and construction material and to grow food, they cleared trees and vegetation from the bluffs to the east and planted every inch of arable land. Because the city's numbers kept increasing, the forest could not return. Instead people kept moving further out to get timber, which then had to be carried considerable distances. Having no beasts of burden, the Cahokians themselves had to do all the carrying. Meanwhile, Woods told me, the city began outstripping its water supply, a "somewhat wimpy" tributary called Canteen Creek. To solve both of these problems at once, the Cahokians apparently changed its course, which had consequences that they cannot have anticipated.

Nowadays Cahokia Creek, which flows from the north, and Canteen Creek, which flows from the east, join together at a point about a quarter mile northeast of Monks Mound. On its way to the Mississippi, the combined river then wanders, quite conveniently, within two hundred yards of the central plaza. Originally, though, the smaller Canteen Creek alone occupied that channel. Cahokia Creek drained into a lake to the northwest, then went straight to the Mississippi, bypassing Cahokia altogether. Sometime between 1100 and 1200 A.D., according to Woods's as-yet unpublished research, Cahokia Creek split in two. One fork continued as before, but the second, larger fork dumped into Canteen Creek. The combined river provided much more water to the city-it was about seventy feet wide. And it also let woodcutters upstream send logs almost to Monks Mound. A natural inference, to Woods's way of thinking, is that the city, in a major public works project, "intentionally diverted" Cahokia Creek.

In summer, heavy rains lash the Mississippi Valley. With the tree cover stripped from the uplands, rainfall would have sluiced faster and heavier into the creeks, increasing the chance of floods and mudslides. Because the now-combined Cahokia and Canteen Creeks carried much more water than had Canteen Creek alone, washouts would have spread more widely across the American Bottom than would have been the case if the rivers had been left alone. Beginning in about 1200 A.D., according to Woods, Cahokia's maize fields repeatedly flooded, destroying the harvests.

The city's problems were not unique. Cahokia's rise coincided with the spread of maize throughout the eastern half of the United States. The Indians who adopted it were setting aside millennia of tradition in favor of a new technology. In the past, they had shaped the landscape mainly with fire; the ax came out only for garden plots of marshelder and little barley. As maize swept in, Indians burned and cleared thousands of acres of land, mainly in river valleys. As in Cahokia, floods and mudslides rewarded them. (How do archaeologists know this? They know it from sudden increases in river sedimentation coupled with the near disappearance of pollen from bottomland trees in those sediments.) Between about 1100 and 1300 A.D., cataclysms afflicted Indian settlements from the Hudson Valley to Florida.

Apparently the majority learned from mistakes; after this time, archaeologists don't see this kind of widespread erosion, though they do see lots and lots of maize. A traveler in 1669 reported that six square miles of maize typically encircled Haudenosaunee villages. This estimate was very roughly corroborated two decades later by the Marquis de Denonville, governor of New France, who destroyed the annual harvest of four adjacent Haudenosaunee villages to deter future attacks. Denonville reported that he had burned 1.2 million bushels of maize-42,000 tons. Today, as I mentioned in Chapter 6, Oaxacan farmers typically plant roughly 1.252.5 acres to harvest a ton of landrace maize. If that relation held true in upstate New York-a big, but not ridiculous assumption-arithmetic suggests that the four villages, closely packed together, were surrounded by between eight and sixteen square miles of maize fields.

Between these fields was the forest, which Indians were subjecting to parallel changes. Sometime in the first millennium A.D., the Indians who had burned undergrowth to facilitate grazing began systematically replanting large belts of woodland, transforming them into orchards for fruit and mast (the general name for hickory nuts, beechnuts, acorns, butternuts, hazelnuts, pecans, walnuts, and chestnuts). Chestnut was especially popular-not the imported European chestnut roasted on Manhattan street corners in the fall, but the smaller, soft-shelled, deeply sweet native American chestnut, now almost extinguished by chestnut blight. In colonial times, as many as one out of every four trees in between southeastern Canada and Georgia was a chestnut-partly the result, it would seem, of Indian burning and planting.

Hickory was another favorite. Rambling through the Southeast in the 1770s, the naturalist William Bartram observed Creek families storing a hundred bushels of hickory nuts at a time. "They pound them to pieces, and then cast them into boiling water, which, after passing through fine strainers, preserves the most oily part of the liquid" to make a thick milk, "as sweet and rich as fresh cream, an ingredient in most of their cookery, especially hominy and corncakes." Years ago a friend and I were served hickory milk in rural Georgia by an eccentric backwoods artist named St. EOM who claimed Creek descent. Despite the unsanitary presentation, the milk was ambrosial-fragrantly nutty, delightfully heavy on the tongue, unlike anything I had encountered before.

Within a few centuries, the Indians of the eastern forest reconfigured much of their landscape from a patchwork game park to a mix of farmland and orchards. Enough forest was left to allow for hunting, but agriculture was an increasing presence. The result was a new "balance of nature."

From today's perspective, the success of the transition is striking. It was so sweeping and ubiquitous that early European visitors marveled at the number of nut and fruit trees and the big clearings with only a dim apprehension that the two might be due to the same human source. One reason that Bartram failed to understand the artificiality of what he saw was that the surgery was almost without scars; the new landscape functioned smoothly, with few of the overreaches that plagued English land management. Few of the overreaches, but not none: Cahokia was a glaring exception.

A friend and I visited Cahokia in 2002. Woods, who lived nearby, kindly agreed to show us around. The site is now a state park with a small museum. From Monks Mound we walked halfway across the southern plaza and then stopped to look back. From the plaza, Woods pointed out, the priests at the summit could not be seen. "There was smoke and noise and sacred activity constantly going on up there, but the peasants didn't know what they were doing." Nonetheless, average Cahokians understood the intent: to assure the city's continued support by celestial forces. "And that justification fell apart with the floods," Woods said.

There is little indication that the Cahokia floods killed anyone, or even led to widespread hunger. Nonetheless, the string of woes provoked a crisis of legitimacy. Unable to muster the commanding vitality of their predecessors, the priestly leadership responded ineffectively, even counterproductively. Even as the flooding increased, it directed the construction of a massive, two-mile-long palisade around the central monuments, complete with bastions, shielded entryways, and (maybe) a catwalk up top. The wall was built in such a brain-frenzied hurry that it cut right through some commoners' houses.

Cahokia being the biggest city around, it seems unlikely that the palisade was needed to deter enemy attack (in the event, none materialized). Instead it was probably created to separate elite from hoi polloi, with the goal of emphasizing the priestly rulers' separate, superior, socially critical connection to the divine. At the same time the palisade was also intended to welcome the citizenry-anyone could freely pass through its dozen or so wide gates. Constructed at enormous cost, this porous architectural folly consumed twenty thousand trees.

More consequential, the elite revamped Monks Mound. By extending a low platform from one side, they created a stage for priests to perform ceremonies in full view of the public. According to Woods's acoustic simulations, every word should have been audible below, lifting the veil of secrecy. It was the Cahokian equivalent of the Reformation, except that the Church imposed it on itself. At the same time, the nobles hedged their bets. Cahokia's rulers tried to bolster their position by building even bigger houses and flaunting even more luxury goods like fancy pottery and jewelry made from exotic semiprecious stones.

It did no good. A catastrophic earthquake razed Cahokia in the beginning of the thirteenth century, knocking down the entire western side of Monks Mound. In 1811 and 1812 the largest earthquakes in U.S. history abruptly lifted or lowered much of the central Mississippi Valley by as much as twelve feet. The Cahokia earthquake, caused by the same fault, was of similar magnitude. It must have splintered many of the city's wood-and-plaster buildings; fallen torches and scattered cooking fires would have ignited the debris, burning down most surviving structures. Water from the rivers, shaken by the quake, would have sloshed onto the land in a mini-tsunami.

Already reeling from the floods, Cahokia never recovered from the earthquake. Its rulers rebuilt Monks Mound, but the poorly engineered patch promptly sagged. Meanwhile the social unrest turned violent; many houses went up in flames. "There was a civil war," Woods said. "Fighting in the streets. The whole polity turned in on itself and tore itself apart."

For all their energy, Cahokia's rulers made a terrible mistake: they did not attempt to fix the problem directly. True, the task would not have been easy. Trees cannot be replaced with a snap of the fingers. Nor could Cahokia Creek readily be reinstalled in its original location. "Once the water starts flowing in the new channel," Woods said, "it is almost impossible to put it back in the old as the new channel rapidly downcuts and establishes itself."

Given Cahokia's engineering expertise, though, solutions were within reach: terracing hillsides, diking rivers, even moving Cahokia. Like all too many dictators, Cahokia's rulers focused on maintaining their hold over the people, paying little attention to external reality. By 1350 A.D. the city was almost empty. Never again would such a large Indian community exist north of Mexico.

HUNDRED YEARS' WAR In the early 1980s I visited Chetumal, a coastal city on the Mexico-Belize border. A magazine had asked me to cover the intellectual ferment caused by the decipherment of Maya hieroglyphics. In researching the article, the photographer Peter Menzel and I became intrigued by the then little-known site of Calakmul, whose existence had been first reported five decades before. Although it was the biggest of all Maya ruins, it had never felt an archaeologist's trowel. Its temples and villas, enveloped in thick tropical forest, were as close to a lost city as we would ever be likely to see. Before visiting Calakmul, though, Peter wanted to photograph it from the air. Chetumal had the nearest airport, which was why we went there.

ABOVE: As late as the 1980s, the Maya city of Kaan (now Calakmul) was encased in vegetation (top). Excavations have now revealed the pyramids beneath the trees (the right-hand mound in the top photo is the pyramid at bottom), exemplifying the recent explosion of knowledge about the Maya.

The town was unpromising in those days. We arrived late in the night, and the only restaurant we could find served a single platter: octopus with pureed beef liver. I am, as a rule, a member of the Clean Plate Club. Looking at the rubbery white octopus chunks bobbing in the tarry mass of liver, I rejected an entire meal for the first time since childhood. Soon afterward the electricity went out everywhere in town. For that reason we did not discover until we retired that our hotel beds were full of little hungry creatures. I was peevish the next morning when we met our pilot.

At first we flew over Highway 186, which arrows west from Chetumal across the Maya heartland. Every so often the pilot tapped my shoulder and pointed to an anonymous, tree-covered hummock. "Ruinas," he said. Otherwise there was little to report. After a while we turned south, toward the border with Guatemala. The Yucatan Peninsula grows wetter as one heads south. The vegetation beneath the plane quickly became thicker, lusher, higher, more aggressive.

All at once we came upon Calakmul. The city proper, built on a low ridge, had once housed as many as fifty thousand people and sprawled across an area as big as twenty-five square miles. (The city-state's total population may have been 575,000.) The downtown area alone had six thousand masonry structures: homes, temples, palaces, and granaries, even an eighteen-foot-high defensive wall. Scattered through the neighborhoods was a network of reservoirs, many apparently stocked with fish. Thousands of acres of farmland extended beyond. Little of this was known then-I am quoting from later reports-and none of it was visible from the plane. From our vantage the only visual testament to Calakmul's past majesty was its two great central pyramids, each wrapped to the shoulders in vegetation.

Peter asked the pilot to fly low circles around the pyramids while he put together the perfect match of light, lens, angle, and shutter speed. He swapped lenses and cameras and window views in a dozen different combinations. At a certain point, he asked, peering through the shutter, "Cuanta gasolina tenemos?" How much gas do we have?

The pilot squinted at the fuel gauge: three-quarters full. A puzzled look spread over his face. I leaned over to watch as he tapped the gauge's foggy plastic cover with his forefinger. The needle plunged almost to empty-it had been pinned.

Peter put down his cameras.

Eventually the blood returned to our heads, permitting cerebration. We had to decide whether to take the shortest path to the airport, straight across the forest, or turn to the north and then fly east along Highway 186, which we could try to land on if we ran out of gas. The trade-off was that the highway route was so much longer that choosing it would greatly increase our chances of a forced touchdown. Soon we realized the decision boiled down to one question: How scary was the prospect of landing in the forest?

I recall looking down at the trees. They had engulfed the great buildings and were slowly ripping apart the soft limestone with their roots. Circling above the city, I had thought, Nobody will ever find out anything about this place. The forest is too overpowering. Calakmul's inhabitants had cut a little divot into its flanks for their city, but now the vegetation, massive and indifferent, was smothering every trace of their existence. From the plane the trees seemed to march to the horizon without interruption.

We flew over the highway. I tried not to stare at the fuel gauge. Still, I couldn't help noticing as, one after another, warning lights blinked on. The plane had so little gas that the engine quit a moment after our wheels hit the tarmac. When we rolled silently to a stop, the pilot leapt out and kissed the ground. I sat back and regarded Chetumal with new affection.

In the mid-1990s the Mexican government paved a road to the site, which is now the center of the 1.7-million-acre Calakmul Biosphere Reserve. Aerial views of the ruins are now spectacular; archaeologists have cleared most of the central city. Along the way, contrary to my initial impression, they have managed to learn a great deal about Calakmul, the landscape it occupied, and the collapse that led to the forest's return. To begin with, they deciphered Calakmul's proper name: Kaan, the Kingdom of the Snake. Impressively, they learned it from the best possible source, the ancient Maya themselves.

Maya scribes wrote in codices made of folded fig-bark paper or deerskin. Unfortunately for posterity, the Spaniards destroyed all but four of these books. The rest of what remains are texts on monuments, murals, and pottery-about fifteen thousand samples of writing, according to one estimate. Piecing together events from these sources is like trying to understand the U.S. Civil War from the plaques on park statues: possible, but tricky. Combining literal interpretation with an understanding of context, epigraphers (decipherers of ancient writing) have spent the last thirty years hauling submerged chunks of Maya history to the surface. David Stuart, a Mayanist at Harvard, decocted the encounter between Chak Tok Ich'aak and the Teotihuacan expedition in 2000. And Simon Martin and Nikolai Grube, respectively of University College London and the University of Bonn, first put the history of the great Mutal-Kaan war together in 1996.

Most of the stelae at Kaan were made from soft stone that has eroded too much to be readable. Martin and Grube thus had to rely on inscriptions at other sites that mention Kaan and its rulers. These are surprisingly numerous. Too numerous, in a way: archaeologists have turned up at least eleven versions of Kaan's early dynastic history painted on big vases. Exasperatingly, none of the eleven tells exactly the same story. The chronological list of rulers differs on different lists, some lists do not include known kings, and some include kings who probably were mythological-as if a tally of English rulers matter-of-factly included King Arthur and his father, Uther Pen-dragon. The dates are inconsistent, too. Kaan's origins may reach back as far as 400 B.C. But the city-state does not unambiguously enter the historical record until about 500 A.D., when it had a king named Yuknoom Ch'een. The city was already dominating its neighbors; in 546, Yuknoom Ch'een's apparent successor supervised the coronation of a five-year-old monarch in nearby Naranjo.

This supervision, recorded on a stela erected seventy years afterward, is the first known example in Yucatan of a Mesoamerican specialty: the chaperoned coronation. For much of the last century most Mayanists believed that at its height-200 to 900 A.D., roughly speaking-the Maya realm was divided into a hugger-mugger of more or less equivalent city-states. Critics pointed out that this theory failed to account for an inconvenient fact: Kaan, Mutal, and a few other cities were much bigger and more imposing than their neighbors, and therefore, one would usually assume, more powerful. According to the skeptics, Maya society was divided into a small number of blocs, each controlled by a dominant city, each striving to achieve some semblance of empire.

Compelling evidence for this view did not emerge until the mid-1980s, when several epigraphers figured out that the Maya glyph ahaw, which means "sovereign" or "lord," had a possessive form, y-ahaw, "his lord," meaning a lord who "belongs" to another lord: that is, a vassal king. Another glyph, u-kahi, turned out to mean "by the action of." They were only two words, but enough to make dozens of texts speak. In the stela, the five-year-old ahaw in Naranjo was crowned "by the action of" the ahaw of Kaan. Naranjo's young king "belonged" to Kaan. (Naranjo is the name scientists gave to the city; its original name may have been Saal.) "The political landscape of the Classic Maya resembles many in the Old World-Classical Greece or Renaissance Italy are worthy comparisons-where a sophisticated and widely shared culture flourished among perpetual division and conflict," Martin and Grube wrote in Chronicles of the Maya Kings and Queens (2000), their remarkable summary of the epigraphic discoveries of the last three decades. It was a "world criss-crossed by numerous patron-client relationships and family ties, in which major centers vied with one another in enmities that could endure for centuries." As Martin put it to me, Maya civilization indeed bore striking similarities to that of ancient Greece. The Greeks were divided into numerous fractious communities, some of which were able to dominate others by threats of force, unequal alliance, or commerce. And just as the conflicted relationship among Athens and Sparta was a leitmotif of Greek life, so Maya society resounded for centuries with the echoes of the struggle between Mutal and Kaan.

Sometime before 561 A.D., a ruler known only as "Sky Witness" took the throne of Kaan. A major figure despite his obscurity, Sky Witness set out to destroy Mutal. The motive for his hatred is uncertain, though it may have been rooted in the invasion from Teotihuacan. The new rulers of Mutal had aggressively thrown their weight around and by Sky Witness's time controlled as much as eight thousand square miles. (Mutal city itself had an estimated population of sixty thousand, plus many more in its hinterland.) Particularly important, the Teotihuacan-backed dynasty took over several outposts on the Usumacinta River system, Yucatan's most important trade route. Shipments of luxury goods from faraway regions usually had to travel up or down the Usumacinta; Mutal's ability to tax and supervise the trade must have been terribly vexing, even if it had little practical import. Sky Witness may have thought that Mutal was becoming a dangerous neighbor and decided to take preemptive action. Or he may have wanted to control the Usumacinta and its tributaries himself. A dynastic dispute may have been involved. Grube told me that he thought the kings of Kaan, never allied with Teotihuacan, may have wanted to stamp out pernicious foreign influences-xenophobia is a powerful motive in every culture. No matter what the motive, Sky Witness's plan to dismantle Mutal was brilliant-in the short run, anyway. In the long run, it helped set in motion the Maya collapse.

Kaan and Mutal had a lot at stake. The Yucatan Peninsula is like a gigantic limestone wharf projecting into the Caribbean. Roughly speaking, the northwest-southeast line on which it joins the mainland runs through the middle of the Maya heartland. Despite receiving three to five feet of precipitation in an average year, this area is prone to drought. Almost all the rain falls during the May-to-December rainy season and rapidly sinks hundreds of feet into the porous limestone, from where it cannot easily be extracted. Little is available during the five hot, dry months between January and April. The region does have permanently water-filled swamps, sinkholes, and lakes, but often these are too salty to drink or use for irrigation. So toxic is the groundwater, a U.S.-Mexican research team remarked in 2002, that the Maya realm was "geochemically hostile" to urban colonization. Its occupation "more resembled settlement on the moon or Antarctica than most other terrestrial habitats."

Most of the salt occurred in the sediments on the swamp bottoms. To make the water potable, the Maya laid a layer of crushed limestone atop the sediments, effectively paving over the salt. As the researchers noted, the work had to be done before the Maya could move in and set up their milpas and gardens. "Permanent, year-round populations could be established only in the presence of an anticipatory engineering of water supplies." The Maya heartland, in other words, was a network of artificially habitable terrestrial islands.

As Maya numbers grew, so did the islands beneath them. North of Kaan, half a dozen small cities improved agricultural conditions by lifting up entire fields and carving out rain-retaining terraces on dry hillsides. Kaan itself dug out a series of reservoirs, established neighborhoods around each one, and linked the ensemble with roads and waterways. Central Mutal was ringed by a chain of seven reservoirs, with another central reservoir reserved for royalty. And so on.

Revamping the landscape both allowed Maya cities to expand and made them more vulnerable. Despite constant maintenance, erosion silted up reservoirs, hurricanes destroyed terraces, and weeds and sediment choked irrigation networks. Over time the Maya found themselves simultaneously maintaining existing systems and pushing out to cover past mistakes. If war damage made it impossible for a city's inhabitants to keep up, they would be in trouble; island dwellers who wreck their homes have no place to move. One can speculate that the losers' fear of having their backs to the wall generated the extraordinary tenacity of the Kaan-Mutal conflict.

Sky Witness's strategy was to ring Mutal with a chain of client states and allies and then strangle it, boa constrictorstyle. In this way Kaan would both acquire a dominant position in the Maya realm and destroy its enemy. The first step was to suborn Mutal's most important vassal, the king of Oxwitza' (now known as Caracol). With 115,000 people, Oxwitza' was twice as populous as Mutal and controlled almost as much territory. Yet it had become Mutal's vassal soon after Teotihuacan installed the new dynasty. No concrete evidence exists that the first event caused the second, but the coincidence in timing is hard to dismiss. Sky Witness seems to have divined or inspired resentment in Oxwitza'. The king of Oxwitza' took the throne "by the action of" Mutal in 553 A.D. Within three years Sky Witness had persuaded the new ruler of Oxwitza' to betray his masters.

Maya polities were not large enough to maintain standing armies; instead both Kaan and Oxwitza' mustered short-term militiamen to fight wars. Wearing cotton armor and wooden helmets, brandishing lances, hatchets, and maces, and carrying great painted litters with effigies of their gods, the two militias marched on Mutal. Kaan is some sixty miles north of Mutal; Oxwitza' is fifty miles south of it. The two cities planned to crush Mutal between them. They carefully chose the day of the attack. Maya priests tracked the movements of Venus, which they regarded as a powerful portent. Its day of emergence in the morning sky was considered an occasion on which warfare and violence was likely to be rewarded-an optimal day to attack a city. On April 29, 562, in what archaeologists call a Star Wars assault, the two celestially guided armies overran Mutal, sacked its precincts, and probably killed its king (the relevant glyphs are too worn to read).

THE HUNDRED YEARS' WAR.

Kaan and Mutal Battle to Control the Maya Heartland, 526682 A.D.

The war between Kaan and Mutal lasted more than a century and consumed much of the Maya heartland. Kaan's strategy was to surround Mutal and its subordinate city-states with a ring of enemies. By conquest, negotiation, and marriage alliance, Kaan succeeded in encircling its enemy-but not in winning the war.

Kaan did not directly occupy Mutal; victorious Maya cities rarely had the manpower to rule their rivals directly. Instead, in the by-now familiar hegemonic pattern, they tried to force the rulers of the vanquished state to become their vassals. If an enemy sovereign was slain, as apparently happened in Mutal, the conquerors often didn't emplace a new one; kings were divine, and thus by definition irreplaceable. Instead the victorious force simply quit the scene, hoping that any remaining problems would disappear in the ensuing chaos. This strategy was partly successful in Mutal-not a single dated monument was erected in the city for a century. Because the city's postattack rulers had (at best) distant connections to the slain legitimate king, they struggled for decades to get on their feet. Unhappily for Kaan, they eventually did it.

The agent for Mutal's return was its king, Nuun Ujol Chaak. Taking the helm of the city in 620 A.D., he was as determined to reestablish his city's former glory as Kaan's rulers were to prevent it. He suborned Kaan's eastern neighbor, Naranjo, which attacked Mutal's former ally, Oxwitza'. The ensuing conflict spiraled out to involve the entire center of the Maya realm. Decades of conflict, including a long civil war in Mutal, led to the formation of two large blocs, one dominated by Mutal, the other by Kaan. As cities within the blocs traded attacks with each other, half a dozen cities ended up in ruins, including Naranjo, Oxwitza', Mutal, and Kaan. The story was not revealed in full until 2001, when a storm uprooted a tree in the ruin of Dos Pilas, a Mutal outpost. In the hole from its root ball archaeologists discovered a set of steps carved with the biography of B'alaj Chan K'awiil, a younger brother or half brother to Mutal's dynasty-restoring king, Nuun Ujol Chaak. As deciphered by the epigrapher Stanley Guenter, the staircase and associated monuments reveal the turbulent life of a great scalawag who spent his life alternately running from the armies of Kaan and Mutal and trying to set them against each other.

Eventually an army under Nuun Ujol Chaak met the forces of Kaan on April 30, 679. Maya battles rarely involved massive direct engagements. This was an exception. In an unusual excursion into the high-flown, the inscriptions on the stairway apostrophized the gore: "the blood was pooled and the skulls of the Mutal people were piled into mountains." B'alaj Chan K'awiil was carried into battle in the guise of the god Ik' Sip, a deity with a black-painted face. In this way he acquired its supernatural power. The technique worked, according to the stairway: "B'alaj Chan K'awiil brought down the spears and shields of Nuun Ujol Chaak." Having killed his brother, the vindicated B'alaj Chan K'awiil took the throne of Mutal.

In a celebration, B'alaj Chan K'awiil joined the rulers of Kaan in a joyous ceremonial dance on May 10, 682. But even as he danced with his master, a counterrevolutionary coup in Mutal placed Nuun Ujol Chaak's son onto the throne, from where he quickly became a problem. On August 5, 695, Kaan soldiers once again went into battle against Mutal. Bright with banners, obsidian blades gleaming in the sun, the army advanced on the long-term enemy it had routed so many times before-and was utterly defeated. In a psychological blow, Mutal captured the effigy of a Kaan patron deity-an enormous supernatural jaguar-that its army carried into battle. A month later, in a mocking ceremonial pageant, the king of Mutal paraded around with the effigy strapped to the back of his palanquin.

Maya "books" consisted of painted bark folded accordion-style into sheaves known as codexes. Time and the Spanish destroyed all but four codexes, and even those are fragmentary (above, a detail from the Paris codex, somewhat reconstructed; right, a piece of the Grolier codex).

Kaan's loss marks the onset of the Maya collapse. Kaan never recovered from its defeat; Mutal lasted another century before it, too, sank into oblivion. Between 800 and 830 A.D., most of the main dynasties fell; cities winked out throughout the Maya heartland. Mutal's last carved inscription dates to 869; soon after, its great public spaces were filled with squatters. In the next hundred years, the population of most southern areas declined by at least three-quarters.

The disaster was as much cultural as demographic; the Maya continued to exist by the million, but their central cities did not. Morley, the Harvard archaeologist, documented the cultural disintegration when he discovered that Maya inscriptions with Long Count dates rise steadily in number from the first known example, at Mutal in 292 A.D.; peak in about 790; and cease altogether in 909. The decline tracks the Maya priesthood's steady loss of the scientific expertise necessary to maintain its complex calendars.

The fall has been laid at the door of overpopulation, overuse of natural resources, and drought. It is true that the Maya were numerous; archaeologists agree that more people lived in the heartland in 800 A.D. than today. And the Maya indeed had stretched the meager productive capacity of their homeland. The evidence for drought is compelling, too. Four independent lines of evidence-ethnohistoric data; correlations between Yucatan rainfall and measured temperatures in western Europe; measurements of oxygen variants in lake sediments associated with drought; and studies of titanium levels in the Caribbean floor, which are linked to rainfall-indicate that the Maya heartland was hit by a severe drought at about the time of its collapse. Water levels receded so deeply, the archaeologist Richardson Gill argued in 2000, that "starvation and thirst" killed millions of Maya. "There was nothing they could do. There was nowhere they could go. Their whole world, as they knew it, was in the throes of a burning, searing, brutal drought.... There was nothing to eat. Their water reservoirs were depleted, and there was nothing to drink."

The image is searing: a profligate human swarm unable to overcome the anger of Nature. Yet Gill's critics dismissed it. Turner, the Clark geographer, was skeptical about the evidence for the killer drought. But even if it happened, he told me, "The entire Maya florescence took place during a prolonged dry period." Because they had spent centuries managing scarce water supplies, he thought it unlikely that they would have fallen quick victim to drought.

Moreover, the collapse did not occur in the pattern one would expect if drought were the cause: in general, the wetter southern cities fell first and hardest. Meanwhile, northern cities like Chichen Itza, Uxmal, and Coba not only survived the dearth of rain, they prospered. In the north, in fact, the areas with the poorest natural endowments and the greatest susceptibility to drought were the most populous and successful. "How and why, then," asked Bruce H. Dahlin, an archaeologist at Howard University, "did the onset of prolonged drought conditions simultaneously produce a disaster in the southern and central lowlands-where one would least expect it-and continued growth and development in the north, again where one would least expect it?"

Dahlin argued in 2002 that Chichen Itza had adapted to the drought by instituting "sweeping economic, military, political, and religious changes." Previous Maya states had been run by all-powerful monarchs who embodied the religion and monopolized trade. Almost all public announcements and ceremonies centered on the figure of the paramount ruler; in the stelae that recount royal deeds, the only other characters are, almost always, the king's family, other kings, and supernatural figures. Beginning in the late ninth or early tenth century A.D., public monuments in Chichen Itza deemphasized the king, changing from official narratives of regal actions to generalized, nontextual images of religion, commerce, and war.

In the new regime, economic power passed to a new class of people: merchants who exchanged salt, chocolate, and cotton from Chichen Itza for a host of goods from elsewhere in Mesoamerica. In previous centuries trade focused on symbolic goods that directly engaged the king, such as jewelry for the royal family. During the drought, something like markets emerged. Dahlins calculated that the evaporation pans outside a coastal satellite of Chichen Itza would have produced at least three thousand tons of salt for export every year; in return, the Maya acquired tons of obsidian for blades, semiprecious stones for jewelry, volcanic ash for tempering pottery, and, most important, maize. Like Japan, which exports consumer electronics and imports beef from the United States and wheat from Australia, Chichen Itza apparently traded its way through the drought.

The contrast between north and south is striking-and instructive. The obvious difference between them was the century and a half of large-scale warfare in the south. Both portions of the Maya realm depended on artificial landscapes that required constant attention. But only in the south did the Maya elite, entranced by visions of its own glory, take its hands off the switch. Drought indeed stressed the system, but the societal disintegration in the south was due not to surpassing inherent ecological limits but the political failure to find solutions. In our day the Soviet Union disintegrated after drought caused a series of bad harvests in the 1970s and 1980s, but nobody argues that climate ended Communist rule. Similarly, one should grant the Maya the dignity of assigning them responsibility for their failures as well as their successes.

Cahokia and the Maya, fire and maize: all exemplify the new view of indigenous impacts on the environment. When scholars first increased their estimates of Indians' ecological management they met with considerable resistance, especially from ecologists and environmentalists. The disagreement, which has ramifying political implications, is encapsulated by Amazonia, the subject to which I will now turn. In recent years a growing number of researchers has argued that Indian societies there had enormous environmental impacts. Like the landscapes of Cahokia and the Maya heartland, some anthropologists say, the great Amazon forest is also a cultural artifact-that is, an artificial object.

Amazonia.

WHAT ORELLANA SAW.

The biggest difficulty in reconstructing the pre-Colombian past is the absence of voices from that past. Mesoamerican peoples left behind texts that are slowly giving up their secrets, but in other areas the lack of written languages has left a great silence. Hints of past events can be found in Native American oral traditions, to be sure, but these are concerned more with interpreting eternal truths than the details of journalism and history. The Bible has much to teach, yet professors must use it judiciously, supplementing it with other sources, when they teach ancient Middle Eastern history. In the same way, preserved Indian lore throws a brilliantly colored but indirect light on the past. To understand long-ago Indian lives, one cannot avoid the accounts of the first literate people who saw them: European swashbucklers, fortune hunters, and missionaries.

As historical sources, colonial reports leave much to be desired. Their authors often were adversaries of the Indians they wrote about, usually did not speak the necessary languages, and almost always had an agenda other than empathetic description of indigenous folkways. Some wrote to further their careers; others, to score political points. Nevertheless these chronicles cannot be dismissed out of hand for these reasons. Used carefully, they can corroborate, even illuminate.

Consider Gaspar de Carvajal, author of the first written description of the Amazon, an account reviled for its inaccuracies and self-serving descriptions almost since the day it was released. Born in about 1500 in the Spanish town of Extremadura, Carvajal joined the Dominican order and went to South America to convert the Inka. He arrived in 1536, four years after Atawallpa's fall. Francisco Pizarro, now governor of Peru, was learning that to avoid outbreaks of feckless violence he needed to keep his men occupied at all times. One of the worst troublemakers was his own half brother, Gonzalo Pizarro. At the time, conquistador society was abuzz with stories of El Dorado, a native king said to possess so much gold that in an annual ritual he painted his body with gold dust and then rinsed off the brilliant coating in a special lake. After centuries of these baths, gold dust carpeted the lake floor. A lake of gold! To twenty-first-century ears the story sounds preposterous, but it did not to Gonzalo Pizarro, who had already helped seize an empire laden with jewels and precious metals. When he decided to search for El Dorado, Francisco encouraged him-he practically shoved Gonzalo out the door. In 1541 Gonzalo left the high Andean city of Quito at the head of an expedition of 200 to 280 Spanish soldiers (accounts differ), 2,000 pigs, and 4,000 highland Indians, the latter slaves in all but name. Accompanying the troops as chaplain was Gaspar de Carvajal.