Although it was just after dawn, several people were already waiting outside the small store. When the metal grating rolled up, I followed them inside. The shop was in a middle-class neighborhood of Oaxaca city, in southern Mexico. Behind the low counter, half a dozen women hovered over waist-high stoves made of concrete block. Recessed into the dome-shaped top of each stove were two shallow clay dishes that served as burners. With expert motions the women slipped tortillas-thin discs of cream-colored flour perhaps nine inches in diameter-onto the hot burners. In seconds the tortilla dried and puffed up like a souffle. And from the storefront floated the aroma of toasting maize, which has permeated Mexico and Central America for thousands of years.

Established in 2001, the tortilla store is an innovative attempt to preserve one of earth's greatest cultural and biological assets: the many local varieties of maize in the narrow "waist" of southern Mexico. The isthmus is a medley of mountains, beaches, wet tropical forests, and dry savannas, and is the most ecologically diverse area in Mesoamerica. "Some parts of Oaxaca go up nine thousand feet," T. Boone Hallberg, a botanist at the Oaxaca Institute of Technology, told me. "Other parts are at sea level. Sometimes the soil is very acid, sometimes it's quite basic-all within a few hundred feet. You can go on either side of a highway, and the climate will be different on the east side than on the west side." The area's human geography is equally diverse: it is the home of more than a dozen major Indian groups, who have a long and fractious history. Despite the strife among them, all of them played a role in the region's greatest achievement, the development of Mesoamerican agriculture, arguably the world's most ecologically savvy form of farming, and of its centerpiece, Zea mays, the crop known to agronomists as maize.

I was visiting Amado Ramirez Leyva, the entrepreneur behind the tortilla store. Born in Oaxaca and trained as an agronomist, Ramirez Leyva had established a consortium of traditional farmers, Indians like himself (Ramirez Leyva is nudzahui [Mixtec], the second most numerous Indian group in the region). The farmers supply eight different varieties of dried maize to his shop, Itanoni, where the kernels are carefully ground, hand-pressed into tortillas, and cooked fresh for customers. Itanoni means "maize flower" in nudzahui, and refers to a flower that blooms in maize fields. It is one of the few tortillerias in Mexico-perhaps even the only one-to sell what might be described as "estate" tortillas: proudly labeled as being made from maize of one variety, from one area.

"Everyone in Mexico knows the rules for making a true tortilla," Ramirez Leyva told me. "But you can't get them that way now, except maybe in your grandmother's kitchen." First soak the dried maize kernels in a bath of lime and water to remove their thin, translucent skins (a process with its own special verb, nixtamalizar). Then stone-grind the kernels into masa, a light, slightly sticky paste with a distinct maize fragrance. Made without salt, spices, leavening, or preservatives, masa must be cooked within a few hours of being ground, and the tortilla should be eaten soon after it is cooked. Hot is best, perhaps folded over with mushrooms or cheese in a tlacoyo. Like a glass of wine, he said, a tortilla should carry the flavor of its native place. "You want to try some?"

A gourmet tortilla shop in Oaxaca, Itanoni is an attempt to preserve southern Mexico's hundreds of varieties of maize, a Mesoamerican tradition that has survived for thousands of years.

I did. The smells in the shop-dry-toasted maize, melting farm cheese, squash flowers sauteing in home-pressed oil-were causing my stomach to direct urgent messages to my brain.

Ramirez Leyva gave me a plateful of tlacoyos. "This is exactly what you would have eaten here ten thousand years ago," he said.

In his enthusiasm, he was overstating, but not by much. Indians didn't have cheese, for one thing. And they didn't eat tortillas ten thousand years ago, though tortillas are indeed ancient. It is known that 11,500 years ago paleo-Indians were hunting from caves in what is now Puebla, the state northwest of Oaxaca. These were not mastodon and mammoth hunters-both species were already extinct. Instead they preyed on deer, horse, antelope, jackrabbit, and, now and then, giant turtle, as well as several species of rodent. Within the next two thousand years all of these animals except deer vanished, too, done in either by a local variant of overkill, the onset of hotter, drier conditions that shrank the available grassland, or both. Responding to the lack of game, people in Oaxaca and Puebla focused more on gathering. Shifting among productive locations, individual families, living on their own, ate seeds and fruit during the spring and fall and hunted during the winter. During the summer they joined together in bands of twenty-five to thirty-cactus leaves, a local favorite, were plentiful enough in that season to support larger groups.

All the while their store of knowledge about the environment increased. People learned how to make agave plants edible (roast them), how to remove the tannic acid from acorns (grind them to a powder, then soak), how to make tongs to pick spiny cactus fruit, how to find wild squash flowers in the undergrowth, and other useful things. Along the way, perhaps, they noticed that seeds thrown in the garbage one year would sprout spontaneously in the next. The sum of these questions led to full-fledged agriculture-not just in the Tehuacan Valley, but in many places in southern Mexico. Squashes, gourds, peppers, and chupandilla plums were among the initial crops. The first cereal was probably millet-not the millet eaten today, which originated in Africa, but a cousin species, knotweed bristle-grass, which is no longer farmed. And then came maize.

At the DNA level, all the major cereals-wheat, rice, maize, millet, barley, and so on-are surprisingly alike. But despite their genetic similarity, maize looks and acts different from the rest. It is like the one redheaded early riser in a family of dark-haired night owls. Left untended, other cereals are capable of propagating themselves. Because maize kernels are wrapped inside a tough husk, human beings must sow the species-it essentially cannot reproduce on its own. The uncultivated ancestors of other cereals resemble their domesticated descendants. People can and do eat their grain; in the Middle East, for example, the wild barley harvest from a small piece of land can feed a family. By contrast, no wild maize ancestor has ever been found, despite decades of search. Maize's closest relative is a mountain grass called teosinte that looks nothing like it (teosinte splits into many thin stems, whereas maize has a single thick stalk). And teosinte, unlike wild wheat and rice, is not a practical food source; its "ears" are scarcely an inch long and consist of seven to twelve hard, woody seeds. An entire ear of teosinte has less nutritional value than a single kernel of modern maize.

The grain in wild grasses develops near the top of the stem. As it matures, the stem slowly breaks up-shatters, in the jargon-letting the seed dribble to the ground. In wild wheat and barley, a common single-gene mutation blocks shattering. For the plant the change is highly disadvantageous, but it facilitates harvest by humans-the grain waits on the stem to be collected. The discovery and planting of nonshattering grain is thought to have precipitated the Neolithic revolution in the Middle East. Like other grasses, teosinte shatters, but there is no known nonshattering variant. (At least sixteen genes control teosinte and maize shattering, a situation so complex that geneticists have effectively thrown up their hands after trying to explain how a nonshattering type might have appeared spontaneously.) No known wild ancestor, no obvious natural way to evolve a nonshattering variant, no way to propagate itself-little wonder that the Mexican National Museum of Culture claimed in a 1982 exhibition that maize "was not domesticated, but created"-almost from scratch.

In the 1960s Richard S. MacNeish, of Phillips Academy, in Andover, Massachusetts, led an archaeological team that meticulously combed Puebla's Tehuacan Valley for signs of early agriculture. Like the Peruvian littoral, the Tehuacan Valley lies in a double rain shadow, sandwiched between two mountain ranges. The aridity similarly helps preserve archaeological evidence. MacNeish's team sifted through fifty caves before they found anything. In site No. 50, a rockshelter near the village of Coxcatlan, the team found maize cobs the size of a cigarette butt.

Ultimately, MacNeish's team found 23,607 whole or partial maize cobs in five caves in the Tehuacan Valley. This ancient refuse became ammunition in a long-running academic battle between Harvard botanist Paul C. Mangelsdorf and George Beadle, a geneticist who worked at Stanford, Caltech, and the University of Chicago. In the late 1930s both men proposed theories about the origin of maize. Mangelsdorf said that it descended from the mix of a now-vanished wild ancestor of maize and wild grasses from the genus Tripsacum. Teosinte, he said, played no role in its development. Beadle had a simpler theory: maize was directly descended from teosinte. Mangelsdorf treated this idea with disbelieving scorn. By now the reader will not be surprised to learn that an apparently arcane debate about the distant past could become vehemently personal. Relations between the two men became cold, then bitter, then explosive. Botanists chose sides and wrote caustic letters about each other.

Mangelsdorf worked with MacNeish and classified the 23,607 ancient maize cobs. The smallest and oldest, he proclaimed, were maize's true wild ancestor, which Indians had then crossed with Tripsacum to make modern maize. So powerful did the evidence of Mangelsdorf's tiny cobs seem that in the 1960s it buried the teosinte hypothesis, even though the latter's champion, Beadle, had for other research won a Nobel Prize. Beadle's ideas were taken up in revamped form by University of Wisconsin botanist Hugh Iltis in 1970. Maize originated, Iltis postulated, in a strange, wholesale mutation of teosinte, to which Indians added and subtracted features through intensive breeding. Mangelsdorf's side found itself on the defensive; Iltis had gleefully pointed out that the "wild maize" cobs from the Tehuacan Valley were identical to those of an unusual, fully domesticated variety of popcorn from Argentina. By then the dispute over the origin of maize had filled almost as much paper-and became as acrimonious-as the battle over Clovis.

In 1997 Mary W. Eubanks, a Duke University biologist, resuscitated the hybridization theory in a new variant. Maize, she suggested, might have been created by repeatedly crossing Zea diploperennis, a rare maize relative, and another cousin species, Eastern gamagrass. When species from different genera hybridize, the result can be what the biologist Barbara McClintock called "genomic shock," a wholesale reordering of DNA in which "new species can arise quite suddenly." In Eubanks's theory, Indians came upon a chance combination of Zea diploperennis and gamagrass and realized that by mixing these two species they could shape an entirely new biological entity. As proof, she announced the creation of a Zea diploperennisgamagrass hybrid in the laboratory that displayed the attributes of ancient maize.

The teosinte faction remained skeptical. A consortium of twelve maize scientists harshly attacked Eubanks's work in 2001 for, in their view, failing to demonstrate that her hybrid was actually a Zea diploperennisgamagrass mix and not an accidental blend of Zea diploperennis and modern maize. (Such errors are a constant threat; in a busy lab, it is all too easy for biologists to use the wrong pollen, mislabel a tray, or mistake one analysis for another.) Meanwhile other geneticists pinpointed teosinte mutations that could have led to modern maize, including sugary 1, a variant gene that alters maize starch in a way that gives tortillas the light, flaky texture celebrated at Itanoni.

Because maize is many steps removed from teosinte, these scientists argued that the modern species had to have been consciously developed by a small group of breeders who hunted through teosinte stands for plants with desired traits. Geneticists from Rutgers University, in New Brunswick, New Jersey, estimated in 1998 that determined, aggressive, knowledgeable plant breeders-which Indians certainly were-might have been able to breed maize in as little as a decade by seeking the right teosinte mutations.

From the historian's point of view, the difference between the two models is unimportant. In both, Indians took the first steps toward modern maize in southern Mexico, probably in the highlands, more than six thousand years ago. Both argue that modern maize was the outcome of a bold act of conscious biological manipulation-"arguably man's first, and perhaps his greatest, feat of genetic engineering," Nina V. Federoff, a geneticist at Pennsylvania State University, wrote in 2003. Federoff's description, which appeared in Science, intrigued me. It makes twenty-first-century scientists sound like pikers, I said when I contacted her. "That's right," she said. "To get corn out of teosinte is so-you couldn't get a grant to do that now, because it would sound so crazy." She added, "Somebody who did that today would get a Nobel Prize! If their lab didn't get shut down by Greenpeace, I mean."

To people accustomed to thinking of maize in terms of dark or light yellow kernels of corn on the cob, the variety in Mexican maize is startling. Red, blue, yellow, orange, black, pink, purple, creamy white, multicolored-the jumble of colors in Mesoamerican maize reflects the region's jumble of cultures and ecological zones. One place may have maize with cobs the size of a baby's hand and little red kernels no bigger than grains of rice that turn into tiny puffs when popped; in another valley will be maize with two-foot-long cobs with great puffy kernels that Mexicans float in soup like croutons. "Every variety has its own special use," Ramirez Leyva explained to me. "This one is for holidays, this one makes tortillas, this one for niquatole [a kind of maize gelatin], this one for tejate," a cold drink in which maize flour, mamey pits, fermented white cacao beans, and other ingredients are marinated in water overnight and then sweetened and whipped to a froth. As a rule domesticated plants are less genetically diverse than wild species, because breeders try to breed out characteristics they don't want. Maize is one of the few farm species that is more diverse than most wild plants.

More than fifty genetically distinguishable maize "landraces" have been identified in Mexico, of which at least thirty are native to Oaxaca, according to Flavio Aragon Cuevas, a maize researcher at the Oaxaca office of the National Institute for Forestry, Agriculture, and Fisheries Research. A landrace is a family of local varieties, each of which may have scores of "cultivars," or cultivated varieties. As many as five thousand cultivars may exist in Mesoamerica.

Maize is open pollinated-it scatters pollen far and wide. (Wheat and rice discreetly pollinate themselves.) Because wind frequently blows pollen from one small maize field onto another, varieties are constantly mixing. "Maize is terribly promiscuous," Hugo Perales, an agronomist at the think tank Ecosur, in Chiapas, told me. Uncontrolled, open pollination would, over time, turn the species into a single, relatively homogeneous entity. But it does not, because farmers carefully sort through the seed they will sow in the next season and generally do not choose obvious hybrids. Thus there is both a steady flow of genes among maize landraces and a force counteracting that flow. "The varieties are not like islands, carefully apart," Perales explained. "They are more like gentle hills in a landscape-you see them, they are clearly present, but you cannot specify precisely where they start."

San Juan Chamula, a mountain town in central Chiapas, near the border with Guatemala, is an example. Located outside the colonial city of San Cristobal de Las Casas, it has a sixteenth-century church with a brilliant blue interior that is a popular tourist destination. But beyond the souvenir kiosks in the cathedral square, most of the 44,000 inhabitants of Chamula scratch a living from the dry mountain slopes outside town. Almost all are Tzotzil, a southwestern branch of the Maya; in 1995, the most recent date for which census data are available, about 28,000 did not speak Spanish. According to a survey by Perales, 85 percent of the farmers plant the same maize landraces as their fathers, varieties that have been passed on and maintained for generations. The crop in the field today is the sum of thousands of individual choices made by community members in the past.

Landrace maize from Oaxaca.

Indian farmers grow maize in what is called a milpa. The term means "maize field," but refers to something considerably more complex. A milpa is a field, usually but not always recently cleared, in which farmers plant a dozen crops at once, including maize, avocados, multiple varieties of squash and bean, melon, tomatoes, chilis, sweet potato, jicama (a tuber), amaranth (a grain-like plant), and mucuna (a tropical legume). In nature, wild beans and squash often grow in the same field as teosinte, the beans using the tall teosinte as a ladder to climb toward the sun; below ground, the beans' nitrogen-fixing roots provide nutrients needed by teosinte. The milpa is an elaboration of this natural situation, unlike ordinary farms, which involve single-crop expanses of a sort rarely observed in unplowed landscapes.

Milpa crops are nutritionally and environmentally complementary. Maize lacks digestible niacin, the amino acids lysine and tryptophan, necessary to make proteins and diets with too much maize can lead to protein deficiency and pellagra, a disease caused by lack of niacin. Beans have both lysine and tryptophan, but not the amino acids cysteine and methionine, which are provided by maize. As a result, beans and maize make a nutritionally complete meal. Squashes, for their part, provide an array of vitamins; avocados, fats. The milpa, in the estimation of H. Garrison Wilkes, a maize researcher at the University of Massachusetts in Boston, "is one of the most successful human inventions ever created."

Wilkes was referring to the ecological worries that beset modern agribusiness. Because agricultural fields are less diverse than natural ecosystems, they cannot perform all their functions. As a result, farm soils can rapidly become exhausted. In Europe and Asia, farmers try to avoid stressing the soil by rotating crops; they may plant wheat one year, legumes the next, and let the field lie fallow in the year following. But in many places this only works for a while, or it is economically unfeasible not to use the land for a year. Then farmers use artificial fertilizer, which at best is expensive, and at worst may inflict long-term damage on the soil. No one knows how long the system can continue. The milpa, by contrast, has a long record of success. "There are places in Mesoamerica that have been continuously cultivated for four thousand years and are still productive," Wilkes told me. "The milpa is the only system that permits that kind of long-term use." Likely the milpa cannot be replicated on an industrial scale. But by studying its essential features, researchers may be able to smooth the rough ecological edges of conventional agriculture. "Mesoamerica still has much to teach us," Wilkes said.

To Wilkes's way of thinking, ancient Indian farming methods may be the cure for some of modern agriculture's ailments. Beginning in the 1950s, scientists developed hybrid strains of wheat, rice, maize, and other crops that were vastly more productive than traditional varieties. The combination of the new crops and the greatly increased use of artificial fertilizer and irrigation led to the well-known Green Revolution. In many ways, the Green Revolution was a tremendous boon; harvests in many poor countries soared so fast that despite burgeoning populations the incidence of hunger fell dramatically. Unfortunately, though, the new hybrids are almost always more vulnerable to disease and insects than older varieties. In addition to being too costly for many small farmers, the fertilizer and irrigation can, if used improperly, damage the soil. Worst, perhaps, in the long run, the exuberant spread of the Green Revolution has pushed many traditional cultivars toward extinction, which in turn reduces the genetic diversity of crops. Wilkes believes that some or all of these difficulties may be resolved by reproducing features of the milpa in a contemporary setting. If this occurs, it will be the second time that the dissemination of Mesoamerican agricultural techniques will have had an enormous cultural impact-the first time being, of course, when they originated.

From today's vantage it is difficult to imagine the impact maize must have had in southern Mexico at the beginning, but perhaps a comparison will help. "Almost pure stands" of einkorn wheat covered "dozens of square kilometers" in Turkey, Iraq, Syria, and other parts of the Middle East, according to Jack R. Harlan and Daniel Zohary, two agronomists who pored over the area in the 1960s to determine the distribution of wild cereals. "Over many thousands of hectares" in those countries, they wrote in Science, "it would be possible to harvest wild wheat today from natural stands almost as dense as a cultivated wheat field." In the Middle East, therefore, the impact of agriculture was thus less a matter of raising the productivity of wheat, barley, and other cereals than of extending the range in which they could be grown, by developing varieties that could flourish in climates and soils that daunt the wild plant. By contrast, the Americas had no wild maize, and thus no wild maize harvest. Stands of teosinte have been seen in the wild, but because the "ears" are tiny and constantly shattering they are difficult to harvest. Thus before agriculture the people of Mesoamerica had never experienced what it was like to stand in a field of grain. Grain fields-landscapes of food!-were part of the mental furniture of people in Mesopotamia. They were an astounding novelty in Mesoamerica. Indians not only created a new species, they created a new environment to put it in. Unsurprisingly, the reverberations sounded for centuries.

Maize in the milpa, the Yale archaeologist Michael D. Coe has written, "is the key...to the understanding of Mesoamerican civilization. Where it flourished, so did high culture." The statement may be more precise than it seems. In the 1970s the geographer Anne Kirkby discovered that Indian farmers in Oaxaca considered it not worth their while to clear and plant a milpa unless it could produce more than about two hundred pounds of grain per acre. Using this figure, Kirkby went back to the ancient cobs excavated from Tehuacan Valley and tried to estimate how much grain per acre they would have yielded. The cob sizes steadily increased as they approached the present. In Kirkby's calculation, the harvest broke the magic two-hundred-pound line sometime between 2000 and 1500 B.C. At about that time, the first evidence of large-scale land clearing for milpas appears in the archaeological record. And with it appeared the Olmec, Mesoamerica's first great civilization.

Based on the Gulf Coast side of Mexico's waist, on the other side of a range of low mountains from Oaxaca, the Olmec clearly understood the profound changes wreaked by maize-indeed, they feted them in their art. Like the stained-glass windows in European cathedrals, the massive Olmec sculptures and bas-reliefs were meant both to dazzle and instruct. A major lesson is the central place of maize, usually represented by a vertical ear with two leaves falling to the side, a talismanic symbol reminiscent of a fleur-de-lys. In sculpture after sculpture, ears of maize spring like thoughts from the skulls of supernatural beings. Olmec portraits of living rulers were often engraved on stelae (long, flat stones mounted vertically in the ground and carved on the face with images and writing). In these stela portraits, the king's clothes, chosen to represent his critical spiritual role in the society's prosperity, generally included a headdress with an ear of maize emblazoned on the front like a star. So resonant was the symbol, according to Virginia M. Fields, curator of pre-Columbian art at the Los Angeles County Museum of Art, that in later Maya hieroglyphics "it became the semantic equivalent of the highest royal title, ahaw." In the Maya creation story, the famous Popul Vuh, humans were literally created from maize.

Maize and the milpa slowly radiated throughout the Americas, stopping their advance only where the climate grew too cold or dry. By the time of the Pilgrims, fields of mixed maize, beans, and squash lined the New England coast and in many places extended for miles into the interior. To the south, maize reached to Peru and Chile. Maize was a high-status food there even though Andean cultures had developed their own agricultural system, with potatoes occupying the central role. (Amazonia seems to have been an exception; most but not all researchers believe maize there was eclipsed by manioc.) Maize had an equivalent impact on much of the rest of the world after Columbus introduced it to Europe. Central Europeans became especially hooked on it; by the nineteenth century, maize was the daily bread of Serbia, Rumania, and Moldavia. So dependent did northern Italy and southwestern France become on polenta, a type of cornmeal mush, that pellagra (caused by eating too much maize) became widespread. "I know little, if anything, pleasing to say about the people," wrote Goethe, who visited northern Italy in 1786. The women's "features indicated misery, and the children were just as pitiful to behold; the men are little better.... The cause of this sickly condition is foundin the continued use of Turkish and heath corn."

Even greater was the impact in Africa, where maize was transforming agriculture by the end of the sixteenth century. "The probability is that the population of Africa was greatly increased because of maize and other American Indian crops," Alfred Crosby told me. "Those extra people helped make the slave trade possible." ("Other American Indian crops" included peanuts and manioc, both now African staples.) Maize swept into Africa as introduced disease was leveling Indian societies. Faced with a labor shortage, the Europeans turned their eyes to Africa. The continent's quarrelsome societies helped them siphon off millions of people. The maize-fed population boom, Crosby believes, let the awful trade continue without pumping the well dry.

THE STUPIDEST QUESTION IN THE WORLD.

A few days after I met Ramirez Leyva, the tortilla entrepreneur, we went to Soledad Aguablanca, a clump of small farms two hours southeast of Oaxaca City. Waiting for us at the side of the road was Hector Diaz Castellano, one of the farmers who supplied Ramirez Leyva's store. Diaz Castellano had a pencil moustache and a rakish straw hat. His Spanish was so heavily salted with Zapotec, the language of Oaxaca's biggest Indian group, that I could not make out a word of it; Ramirez Leyva had to translate. The maize field was at the end of a long, rutted dirt road that led up a rise. Although we had left just after dawn, the sun was hot enough by our arrival to make me wish for a hat. Diaz Castellano walked along the rows, his gaze taking in every stalk as he passed. For an hour he spoke, almost without stopping, about his maize and the market for his maize. He was not, I suspected, a naturally loquacious man, but that morning he had a subject that interested him.

Hector Diaz Castellano.

Diaz Castellano's maize field was one of the 340,000 farms in Oaxaca. His farm, like about two-thirds of the farms in the state, occupied less than ten acres-unviably small by the standards of developed nations. Most landrace maize is grown on these farms, partly because of tradition and partly because they are usually in areas that are too high, dry, steep, or exhausted to support high-yield varieties (or owned by farmers too poor to afford the necessary fertilizer). As if being grown on tiny farms in bad conditions weren't enough, landrace maize is usually less productive than modern hybrids; a typical yield is .4 to .8 tons per acre, whereas Green Revolution varieties in Oaxaca reap between 1.2 and 2.5 tons per acre when properly fertilized, a crippling advantage. The meager harvests may be enough for subsistence but can rarely be brought to market because farm villages are often hours away on dirt roads from the nearest large town. But even when farmers try, it is often little use: modern hybrids are so productive that despite the distances involved U.S. corporations can sell maize for less in Oaxaca than can Diaz Castellano. Landrace maize, he said, tastes better, but it is hard to find a way to make the quality pay off. He was lucky, he said, that Ramirez Leyva was trying to market his crop.

We went to Diaz Castellano's house for breakfast. His wife, Angelina, round and short-haired in a tight plaid dress, was cooking tortillas in an outdoor shed with corrugated aluminum walls. A wood fire burned beneath a concave clay griddle called a comal. The comal was propped above the flames on three rocks-a cooking method as old as Mesoamerican culture. By the fire, in a three-legged stone bowl, was a lump of fresh masa twice the size of a toaster. The stereotype is that rural Mexicans are generous to strangers. Piling my plate high, Angelina did nothing to dispel this impression.

I asked her husband what he was. I had wanted to find out which Indian group he was born into, but he took the question another way.

"Somos hombres de maiz," he said, enunciating clearly for my benefit. We are men of maize.

I wasn't sure what to make of this gnomic utterance. Was he pulling my leg?

"Everybody says that," Ramirez Leyva said, observing my confusion. "It's an idiom." A little while later I visited a Danish anthropologist at the International Maize and Wheat Improvement Center (CIMMYT), outside Mexico City. Watching films of her interviews with Oaxacans, I saw two old women explain to the young anthropologist that they, too, were hombres de maiz. So Ramirez Leyva was right, I thought. A day later a CIMMYT biologist gave me a paperback book, describing it as "the best novel ever written about Mesoamerica." It was Hombres de maiz, by Miguel Angel Asturias. All right already, I thought. I get it.

Meanwhile Angelina had come out from behind the comal and joined her husband. In the Oaxacan countryside, they explained to me, a house without maize growing in the backyard is like a house without a roof or walls. You would never not have maize, they said. They were speaking matter-of-factly, as if telling me how to take the bus. Even in the city, they said, where people cannot grow maize, nobody would even think of passing a day without eating it.

Curious, I asked what they thought would happen if they didn't have maize every day. Diaz Castellano looked at me as if I had asked the stupidest question in the world.

"Why should I want to be somebody else?" he said.

Writing, Wheels, and Bucket Brigades.

(Tales of Two Civilizations, Part II).

"LIKE GRAPES THEY FALL OFF"

On January 16, 1939, Matthew W. Stirling took an early-morning walk through the wet, buggy forest of Veracruz state, on the Gulf Coast side of Mexico's southern isthmus. Eighty years before his walk, a villager traipsing through the same woods had stumbled across a buried, six-foot-tall stone sculpture of a human head. Although the find was of obvious archaeological importance, the object was so big and heavy that in the intervening eight decades it had never been pulled out of the ground. Stirling, director of the Smithsonian Bureau of American Ethnology, had gone to Mexico the year before, in early 1938, to see the head for himself. He found it, sunk to the eyebrows in mud, after an eight-hour horseback ride from the nearest town. The head was in the midst of about fifty large, artificial earthen mounds-the ruins, Sterling concluded with excitement, of a previously unknown Maya civic center. He had decided to assemble a research team and explore the area in more detail the next year, and persuaded the National Geographic Society to foot the bill. When he returned to Veracruz, he and his team cleared the dirt around the great head, admiring its fine, naturalistic workmanship, so unlike the stiff, stylized sculpture common elsewhere in Mesoamerica. Nearby, they found a stela, its wide, flat face covered with bas-relief figures. Hoping to turn up others, Stirling was walking that January morning to the far end of the mounded area, where a workman had noticed a large, flat, partly submerged rock: a second stela.

Accompanying him were twelve workers from the nearby hamlet of Tres Zapotes. They pried the stela from the ground with wooden poles, but it was blank. Disappointed, Stirling took the crew to yet a third fallen stela. They scraped away the covering dirt and found that it, like the first, was covered with intricate images. Alas, the carvings were now too weathered to be deciphered. The frustrated Stirling asked the workers to expose the back of the slab by digging beneath it and levering up the stone with poles. Several of the men, he later recounted, "were on their knees in the excavation, cleaning the mud from the stone with their hands, when one of them spoke up in Spanish: 'Chief! Here are numbers!'"

Across the back of the stela were clumps of dots and bars, a notation familiar to Stirling from the Maya. The Maya used a dot to signify one and a horizontal bar to signify five; the number nineteen would thus be three bars and four dots. Stirling copied the dots and bars and "hurried back to camp, where we settled down to decipher them." The inscription turned out to be a date: September 3, 32 B.C, in today's calendar.

Stirling already knew that Tres Zapotes was anomalous-it was at least 150 miles west of any previously discovered Maya settlement. The date deepened the puzzle. If, as seemed likely, it recorded when the stela was put on display, this implied that Tres Zapotes had been a going concern in 32 B.C.-centuries before any other known Maya site. The date thus seemed to imply that the Maya had originated well to the west of what was thought of as their traditional homeland, and much earlier than had been thought. Stirling didn't believe it. Surely the Maya had not sprung up in Tres Zapotes and then moved en masse hundreds of miles to the east. But the alternative explanation-that Tres Zapotes was not a Maya community-seemed equally improbable. The Maya were universally regarded as the oldest advanced society in Mesoamerica. Whoever had carved the stela had some knowledge of writing and mathematics. If they were not Maya, the implication was that someone else had launched the project of civilization in Mesoamerica.

Learning from local people that Tres Zapotes was only one of many mound sites in Veracruz, Stirling decided to return in 1940 to survey them all. The task was daunting even for a cigar-chomping, whisky-drinking, adventure addict like Stirling. Most of the mound centers were in the middle of trackless mangrove swamps or up narrow, unmapped rivers choked with water hyacinth. Ticks and mosquitoes were indefatigable and present in huge numbers; the ticks were worse than the mosquitoes, Stirling remarked, because they had to be dug out of the flesh with a knife. At one point Stirling and a colleague hitched a ride in a pepper truck to one of the smaller sites. After jolting down a road with deep ruts "designed to test the very souls of motorcars," the two men were let off in a nondescript meadow. Stirling went to talk with the driver.

"The ticks are not bad, are they?" I asked him hopefully, viewing the tall grass and underbrush between the road and the mounds. "No," said the driver, beaming. "When full, like grapes they fall off and no harm is done. There are millions of them here, however."

In La Venta, a dry, raised "island" in the coastal swamp, Stirling's team discovered four more colossal heads. Like the first, they had no necks or bodies and wore helmets that vaguely resembled athletic gear. All were at least six feet tall and fifteen feet round and made from single blocks of volcanic basalt. How, Stirling wondered, had their makers transported these ten-ton blocks from the mountains and across the swamp? Whoever these people were, he eventually concluded, they could not be Maya; their ways of life seemed too different. Instead they must have belonged to another culture altogether. La Venta was filled with mounds and terraces, which told Stirling that many people had lived there. The city, he wrote in 1940, "may well be the basic civilization out of which developed such high art centers as those of the Maya, Zapotecs, Toltecs, and Totonacs." He called its "mysterious people" the Olmec.

Stirling's account set the template for decades to follow. Ever since his day, the Olmec have been known by two Homeric epithets: they were "mysterious," and they were the "mother culture" of Mesoamerica. (Tourists are told by Frommer's 2005 Mexico guide, for example, to visit the ruins of the "enigmatic people" who created the "mother culture of Mesoamerica.") But in recent years many archaeologists have come to believe that neither description is correct.

Curious villagers surround the great Olmec head excavated in 1939 by archaeologist Matthew Stirling in the Mexican state of Veracruz.

The Olmec's purported mysteriousness is related to their emergence. To Stirling and many of his successors, the Olmec seemed to have no peers or ancestors; they appeared fully formed, apparently from nowhere, like Athena springing from the brow of Zeus. First there was a jungle with a few indistinguishable villages; then, suddenly, a sophisticated empire with monumental architecture, carved stelae, earthwork pyramids, hieroglyphic writing, ball courts, and fine artworks-all of it conjured into existence with the suddenness of amagician's trick. The Olmec, wrote Smithsonian archaeologist Betty Meggers, were a "quantum change." Their status as precursors led archaeologists to believe that the subsequent emergence of other complex societies was due to their example-or their conquest. Even the mighty Maya did little more than continue down the path set by the Olmec. "There is now little doubt," Yale archaeologist Michael Coe wrote in 1994, "that all later civilizations in Mesoamerica, whether Mexican or Maya, ultimately rest on an Olmec base."

Strictly speaking, Coe was mistaken. By the time he wrote, many of his colleagues strongly doubted that the Olmec either emerged alone or were the mother culture. They did emerge abruptly, these researchers say, but they were only the first of the half-dozen complex societies-"sister cultures"-that sprang up in southern Mexico after the development of maize agriculture. Focusing on the Olmec's chronological primacy, they believe, obscures the more important fact that Mesoamerica was the home of a remarkable multisociety ferment of social, aesthetic, and technical innovation.

RUBBER PEOPLE.

Nobody knows the right name for the Olmec, but "Olmec" is the wrong one. They spoke a language in the Mixe-Zoquean language family, some members of which are still used in isolated pockets of southern Mexico. "Olmec," though, is a word in Nahuatl, the language of the Mexica to the north. It means, more or less, "people of the land of rubber." The problem with the name is not so much that the Olmec did not use it for themselves-nobody knows what that name was, and they have to be called something. Nor is the problem the rubber, which the Olmec used, and may have invented (scientists discovered in the 1990s that they made rubber by chemically treating the latex-containing sap of a tropical tree, Castilla elastica). The problem is that the Mexica did not actually use the name to refer to the putative mother culture in Veracruz, but to another, completely unrelated culture in Puebla to the west, a culture that, unlike the ancient Olmec, still existed at the time of the Spanish conquest. The confusion between the Mexica's Olmec and Stirling's Olmec led some archaeologists to propose that the latter should be called the "La Venta Culture," after the site he investigated. Almost everyone agreed that the new name was a big improvement, logically speaking. Unfortunately, nobody used it. Not for the first time in Native American history, the confusing, incorrect name prevailed.

The Olmec heartland was the coastal forests of Veracruz. Compared to the Norte Chico, the area is promising. Like the Peruvian littoral, it is bracketed by sea and mountains, but it catches, rather than misses, the prevailing winds, and the rain that comes with them. The shoreline itself is swampy, but not far from the coast the country rises into a lush, fertile plateau. Further inland are the Tuxtla Mountains, with many rivers cascading down their flanks. The rivers flood in the rainy season, enriching the land, Nile Delta style. During the rest of the year, the climate is drier, and farmers plant and tend their milpas on the alluvial soil.

The first traces of the people who would become the Olmec date back to about 1800 B.C. At that time there was little to distinguish them from groups elsewhere in Mesoamerica. But something happened in Veracruz, some spark or incitement, a cultural quickening, because within the next three centuries the Olmec had built and occupied San Lorenzo, the first large-scale settlement in North America-it covered 2.7 square miles. On a plateau commanding the Coatzacoalcos river basin, San Lorenzo proper was inhabited mainly by the elite; everyone else lived in the farm villages around it. The ceremonial center of the city-a series of courtyards and low mounds, the latter probably topped with thatch houses-sat on a raised platform 150 feet high and two-thirds of a mile to a side. The platform was built of almost three million cubic yards of rock, much of it transported from mountain quarries fifty miles away.

Scattered around the San Lorenzo platform were stone monuments: massive thrones for living kings, huge stone heads for dead ones. Rulers helped to mediate between supernatural forces in the air above and the watery place below where souls went after life. When kings died, their thrones were sometimes transformed into memorials for their occupants: the colossal heads. The features of these enormous portraits are naturalistically carved and amazingly expressive-thoughtful or fiercely proud, mirthful or dismayed. It is assumed they were placed like so many stone sentinels for maximum Orwellian impact: the king is here, the king is watching you.*20 Like the carvings and stained-glass windows in European cathedrals, the art in San Lorenzo and other Olmec cities consisted mainly of powerful, recurring images-the crucifixions and virgins, so to speak, of ancient Mesoamerica. Among these repeated subjects is a crouched, blobby figure with a monstrously swollen head. Puzzled researchers long described these sculptures as "dwarves" or "dancers." In 1997 an archaeologist and a medical doctor with archaeological leanings identified them as human fetuses. Their features were portrayed accurately enough to identify their stage of development. Researchers had not recognized them because artistic renditions of fetuses are almost unheard of in European cultures (the first known drawing of one is by Leonardo). Other frequent themes included lepers, the pathologically obese, and people with thyroid deficiencies, all portrayed with a cool eye for anatomical detail. Perhaps the best-known subject is a man or boy gingerly holding a "were-jaguar": a limp, fat, sexless baby with a flattened nose and a snarling jaguar mouth. Often the baby has a deeply cloven skull.

The denizens of San Lorenzo are unlikely to have shared Europeans' dismay at the physical deformity portrayed in these images. Indeed, by contemporary standards high-born Olmec were deformed themselves. By binding small, flat pieces of wood to newborns' foreheads, they pushed up the soft infant bones, making the skull longer and higher than normal. To further proclaim their status, wealthy Olmec carved deep grooves into their teeth and pierced their nasal septums with bone awls, plugging the holes with ornamental jade beads. (Because no Olmec skeletons have been found, no direct proof of these practices exists; instead archaeologists base their beliefs on the portrayal of Olmec nobles in figurines and sculptures.) Swanning about the elite precincts, the rich and powerful wore finely woven clothing, but only below the waist-breechclouts for men, skirts and belts for women. Veracruz was too hot for anything more. On public occasions, nobles bedizened themselves with bracelets, anklets, many-stranded necklaces, bejeweled turbans, and big, hiphop-style pendants. Some of the last were concave mirrors made from beautifully polished magnetite. Precisely ground, the mirrors were able to start fires and project images onto flat surfaces, camera lucida fashion. Presumably they were used to dazzle hoi polloi. As for the poor, it is likely that they went naked, except possibly for sandals.

San Lorenzo fell in about 1200 B.C., victim of either revolution or invasion. Or perhaps it was abandoned and sacked for religious reasons-archaeologists have advanced several hypotheses for the city's demise. What is certain is that the site was vacated and the stelae defaced and the sculptures decapitated. The colossal heads being, so to speak, pre-decapitated, they were smashed with hammers and systematically buried in long lines. Vegetation overran the red-ocher floors and the workshops that manufactured ceramic figurines and iron beads and rubber ax-head straps.

Olmec society was surprisingly unaffected by the collapse of its greatest polity. A much bigger city, La Venta, was going up on a swamp island about forty miles away.

Today La Venta is partly buried by an oil refinery, but in its heyday-roughly speaking, 1150 B.C. to 500 B.C.-it was a large community with a ring of housing that surrounded a grand ceremonial center. The city's focus, its Eiffel Tower or Tiananmen Square, was a 103-foot-tall clay mound, a bulging, vertically fluted cone somewhat resembling a head of garlic. The mound rose at the south end of a rectangular, hundred-yard-long pavilion that was bordered by two knee-high berms. At the north end of the pavilion was a sunken rectangular courtyard fenced on three sides by a row of seven-foot basalt columns atop a low red and yellow adobe wall. The fourth, northern side opened onto a third mound, larger than the small mound but nowhere near the size of the big one. The pavilion and courtyard had painted walls and floors of colored sand and clay; heavy sculptural objects, including several of the trademark heads, studded the area. This central part of the city was reserved, archaeologists believe, for clerics and rulers. It was Buckingham Palace and the Vatican rolled into one.

La Venta, too, was destroyed, perhaps deliberately, around 350 B.C. But its eight-hundred-year existence spanned one of the most exciting times in American history. At La Venta's height, Olmec art and technical innovations could be found throughout Mesoamerica. So widespread was Olmec iconography-jaguar babies, carved stelae, distinctively shaped ceramics-that many archaeologists believed its very ubiquity was evidence that the Olmec "not only engendered Mesoamerica but also brought forth the first Mesoamerican empire." The description is from Ignacio Bernal, once director of Mexico's famed National Museum of Anthropology. Bernal, who died in 1992, envisioned an imperium that spanned much of southern Mexico, proselytizing its religion and forcing other groups to send their finest works to its heartland. The Olmec, he thought, were the Romans of Mesoamerica, a magisterial society that "established the pattern which, through the centuries, was to be followed by other expansionist Mesoamerican cultures."

Since Bernal's death many researchers have come to view the Olmec differently. According to the University of Michigan anthropologists Kent Flannery and Joyce Marcus, the Olmec heartland was but one of four regional power centers: the Central Basin to the north, where settlements like Tlatilco and Tlapacoya laid the groundwork for the empires of Teotihuacan and the Toltecs; in the isthmus, the chiefdoms of Oaxaca; the Olmec, along the Gulf Coast; and, later, the Maya polities in Yucatan and northern Guatemala. Some believe that there was a fifth power: Chalcatzingo, an important chiefdom between the Central Basin and Oaxaca.

In the first millennium B.C., all four (or five) were making the transition from individual fortified villages to groups of chiefdoms to states with centralized authority. Although the Olmec preceded the others, they did not set the template for them. Instead they all influenced each other, sometimes by trade, sometimes by violence, each one developing new techniques, exporting unique goods, and swiping ideas from the others. In this world of "competitive interaction," all parties hustled for advantage. Trade in goods was important, but it was the trade in ideas that mattered.

By making this argument, I am endorsing one side in the long-running dispute between mother- and sister-culture proponents. In other words, as one mother-culture advocate put it, I am "swallowing Marcus's [nonsense] whole." He may be right. But I would argue that there is a difference between inheriting a cultural tradition, as the Norte Chico culture's successors apparently did, and copying one, as the Olmec's proponents argue. Nobody disputes that Han Chinese society arose long before its neighbors in Asia, but Asian archaeologists don't refer to China as Asia's "mother culture," because China's neighbors used part of its intellectual legacy to build up their own distinct and different writing systems, agricultural technologies, imperial practices, and much else. Given the evidence available now, the same seems to be true for the Olmec and ancient Mesoamerica. (To me, anyway-many researchers disagree.) MESOAMERICA, 1000 B.C.1000 A.D.

For thousands of years Mesoamerica was a wellspring of cultural innovation and growth. This map does not depict it accurately, because these societies were not contemporaries-the Olmec vanished centuries before the nudzahui and Zapotec began to reach their height, for example. Nonetheless, it should give some idea of each society's heartland.

Emblematic of this rocketing growth were the Olmec's neighbors in Oaxaca, the Zapotec, whom Flannery and Marcus have studied for more than two decades. The Zapotec were based across the mountains from the Olmec, in Oaxaca's high Central Valley-three forty-to-sixty-mile-long bowls that intersect in a ragged Y. By about 1550 B.C., they were abandoning the life of hunting and gathering to live in villages with defensive palisades.*21 These early villages had wattle-and-daub houses, fine pottery, and some public architecture. They were controlled by "big men," the social scientist's term for the alpha male who is able in such informal settings to enforce his will through persuasion or force. Within a few hundred years, the big men acquired rank-that is, they began to wield power not only because of their personal charisma, but also because their societies had given them an elevated official position. It is the difference between the leading spirit on a pickup basketball team and the officially selected captain of a college squad. After this, political consolidation proceeded apace. Soon the valley was dominated by three main chiefdoms, one at each endpoint of the Y. They did not get along; a thirty-square-mile buffer zone in the middle, Flannery and Marcus noted, was "virtually unoccupied."

The biggest of the three chiefdoms was located near today's village of San Jose Mogote. Around 750 B.C. it was attacked and its temple razed in a fire so intense that the clay walls melted. San Jose Mogote quickly rebuilt the temple a few yards away.

Across the new threshold artisans laid a carved stone. Splayed across its face, suitable for stomping on, was a bas-relief of a naked, disemboweled male corpse, apparently a defeated enemy, blood welling from his side. The carving, to my taste, is beautifully done, as graphically elegant, despite the gory subject matter, as a Matisse. What matters most to researchers, though, is not the design but the two curious marks between the cadaver's feet. The two marks are glyphs: the oldest firmly dated writing in the Americas. Despite their brevity, they are full of information, both because of the nature of written language, and because of the special way Mesoamerican people chose their names.

COUNTING AND WRITING.

Writing begins with counting. When a culture grows big enough, it acquires an elite, which needs to monitor things it considers important: money, stored goods, births and deaths, the progression of time. In the Fertile Crescent, village accountants began keeping records with clay tokens around 8000 B.C. As the need for precision grew, they scratched marks on the tokens as mnemonic devices. For example, they might have distinguished a count of sheep from one of wheat by drawing a sheep on one and a wheat stalk on the other. Gradually the information on each record increased. The bureaucrats were not intending to create writing. Instead they were simply adding useful features as they became necessary. By 3200 B.C. Sumerian scribes had progressed to inscribing on clay tablets with sharpened reeds. A tablet might contain, say, two hash marks, a box, a circle with a cross in the middle, an asterisk-like shape, and an arrangement of three triangles. Scribes would know that the hash marks meant "two," the box was a "temple," the circle stood for "cattle," an asterisk meant "goddess," and the triangles were "Inanna"-two cattle owned by the goddess Inanna's temple. (Here I am lifting an example from Gary Urton, a Harvard anthropologist.) They had no way to indicate verbs or adjectives, no way to distinguish subject from object, and only a limited vocabulary. Nonetheless, Sumerians were moving toward something like writing.

In Mesoamerica, timekeeping provided the stimulus that accounting gave to the Middle East. Like contemporary astrologers, the Olmec, Maya, and Zapotec believed that celestial phenomena like the phases of the moon and Venus affect daily life. To measure and predict these portents requires careful sky watching and a calendar. Strikingly, Mesoamerican societies developed three calendars: a 365-day secular calendar like the contemporary calendar; a 260-day sacred calendar that was like no other calendar on earth; and the equally unique Long Count, a one-by-one tally of the days since a fixed starting point thousands of years ago. Establishing these three calendars required advances in astronomy; synchronizing them required ventures into mathematics.

The 260-day ritual calendar may have been linked to the orbit of Venus; the 365-day calendar, of course, tracked the earth's orbit around the sun. Dates were typically given in both notations. For example, October 12, 2004, is 2 Lamat 11 Yax, where 2 Lamat is the date in the ritual calendar and 11 Yax the date in the secular calendar. Because the two calendars do not have the same number of days, they are not synchronized; the next time 2 Lamat occurs in the sacred calendar, it will be paired with a different day in the secular calendar. After October 12, 2004, in fact, 2 Lamat and 11 Yax will not coincide again for another 18,980 days, about fifty-two years.

Mesoamerican cultures understood all this, and realized that by citing dates with both calendars they were able to identify every day in this fifty-two-year period uniquely. What they couldn't do was distinguish one fifty-two-year period from another. It was as if the Christian calendar referred to the year only as, say, '04-one would then be unable to distinguish between 1904, 2004, and 2104. To prevent confusion, Mesoamerican societies created the third calendar, the Long Count. The Long Count tracks time from a starting point, much as the Christian calendar begins with the purported birth date of Christ. The starting point is generally calculated to have been August 13, 3114, B.C., though some archaeologists put the proper date at August 10 or 11, or even September 6. Either way, Long Count dates consisted of the number of days, 20-day "months," 360-day "years," 7,200-day "decades," and 144,000-day "millennia" since the starting point. Archaeologists generally render these as a series of five numbers separated by dots, in the manner of Internet Protocol addresses. Using the August 13 starting date, October 12, 2006, would be written in the Long Count as 12.19.13.12.18. (For a more complete explanation, see Appendix D.) Because it runs directly from 1 B.C. to 1 A.D., the Christian calendar was long a headache for astronomers. Scientists tracking supernovae, cometary orbits, and other celestial phenomena would still have to add or subtract a year manually when they crossed the A.D.-B.C. barrier if a sixteenth-century astronomer named Joseph Scaliger hadn't got sick of the whole business and devised a calendar for astronomers that doesn't skip a year. The Julian calendar, which Scaliger named after his father, counts the days since Day 0. Scaliger chose Day 0 as January 1, 4713, B.C.; Day 1 was January 2. In this system, October 12, 2006, is Julian Day 2,454,021.

The Long Count calendar began with the date 0.0.0.0.0.*22 Mathematically, what is most striking about this date is that the zeroes are true zeroes. Zero has two functions. It is a number, manipulated like other numbers, which means that it is differentiated from nothing. And it is a placeholder in a positional notation system, such as our base-10 system, in which a number like 1 can signify a single unit if it is in the digits column or ten units if it is in the adjacent column.

That zero is not the same as nothing is a concept that baffled Europeans as late as the Renaissance. How can you calculate with nothing? they asked. Fearing that Hindu-Arabic numerals-the 0 through 9 used today-would promote confusion and fraud, some European authorities banned them until the fourteenth century. A classic demonstration of zero's status as a number, according to science historian Dick Teresi, is grade point average: In a four-point system, an A equals 4, B equals 3, and so on, down to E, which equals 0. If a student takes four courses and gets A's in two but fails the other two, he receives a GPA of 2.0, or a C average. The two zeroes drag down the two A's. If zero were nothing, the student could claim that the grades for the courses he failed did not exist, and demand a 4.0 average. His dean would laugh at such logic.

Without a positional notation system, arithmetic is tedious and hard, as schoolchildren learn when teachers force them to multiply or subtract with Roman numerals. In Roman numerals, CLIV is 154, whereas XLII is 42. Maddeningly, both numbers have L (50) as the second symbol, but the two L's aren't equivalent, because the second is modified by the preceding X, which subtracts ten from it to make forty. Even though both CLIV and XLII are four-digit numbers, the left-hand symbol in the first number (C) cannot be directly compared with the left-hand symbol in the second (X). Positional notation symbols take the aggravation out of arithmetic.

Stirling's stela in Tres Zapotes bore a Long Count date of 7.16.6.16.18. The implication is that by 32 B.C. the Olmec already had all three calendars and zero to boot. One can't be sure, because the date does not include a zero or a reference to the other calendars. But it is hard to imagine how one could have a Long Count without them. Tentatively, therefore, archaeologists assign the invention of zero to sometime before 32 B.C., centuries ahead of its invention in India.

How long before 32 B.C.? The carved cadaver in San Jose Mogote may give a hint. In Mesoamerican cultures, the date of one's birth was such an important augury of the future that people often acquired that day as their name. It was as if coming into the world on New Year's Day were such a sign of good fortune that children born on that day would be named "January 1." This seems to have been the case for the man whose death was celebrated in the San Jose Mogote temple. Between his feet are two glyphs, one resembling a stovepipe hat with a U painted across the front, the other looking vaguely like a smiling pet monster from a Japanese cartoon. According to Marcus, the Michigan anthropologist, the glyphs correspond to 1-Earthquake, the Zapotec name for the seventeenth day of the 260-day sacred calendar. Because the carving depicts a man instead of an event, the date is generally thought to be the dead man's name. If so, 1-Earthquake is the first named person in the history of the Americas. Even if the date is not a name, the two glyphs indicate that by 750 B.C., when the slab was carved, the Zapotec were not only on the way to some form of writing, but had also assembled some of the astronomical and mathematical knowledge necessary for a calendar.

Discovered in 1975, this prone, disemboweled man was carved onto the stone threshold of a temple in San Jose Mogote, near the city of Oaxaca. Between the corpse's feet is the oldest certainly dated writing in the Americas: two glyphs (shaded in drawing) that probably represent his name, 1-Earthquake. The ornate scroll issuing from his side is blood. According to Joyce Marcus, the first archaeologist to examine this bas-relief, the Zapotec words for "flower" and "sacrificial object" are similar enough that the flowery blood may be a visual pun.

To judge by the archaeological record, this development took place in an astonishingly compressed period; what took the Sumerians six thousand years apparently occurred in Mesoamerica in fewer than a thousand. Indeed, Mesoamerican societies during that time created more than a dozen systems of writing, some of which are known only from a single brief text. The exact chronology of their evolution remains unknown, but could be resolved by the next object that a farmer discovers in a field. The earliest known Olmec writing, for example, is on a potsherd from Chiapas that dates from about 300 B.C. For a long time nobody could read it. In 1986 a workcrew building a dock on the Acula River in Veracruz pulled out a seven-foot stela covered with Olmec symbols. Thought to have been written in 159 A.D., the twenty-one columns of glyphs were the first Olmec text long enough to permit linguists to decipher the language. Two linguists did just that in 1993. The stela recounted the rise of a warrior-king named Harvest Mountain Lord who celebrated his ascension to the throne by decapitating his main rival during the coronation. This information in hand, the linguists went back to the writing on the potsherd. Disappointingly, it turned out to be some banal utterances about dying and cutting cloth.

FROM 1-EARTHQUAKE TO 8-DEER.

The development of writing in Zapotec society went hand-in-hand with growing urbanization. In about 500 B.C., San Jose Mogote seems to have transplanted itself to Monte Alban, in the middle of the buffer zone. About half an hour by bus from Oaxaca City, Monte Alban is today a decorous sprawl of walls and pyramids enveloped by a lush lawn (this last is an import from Europe; lawn grass did not exist in the Americas prior to Columbus). Arriving tourists are hailed by "guides" with backpacks full of phony ancient figurines and ethnically incorrect souvenirs of Mexica drawings. Their ministrations do not diminish the lonely dignity of the ruins. Monte Alban is atop a steep, 1,500-foot hill that overlooks the valley of Oaxaca. The Zapotec reconfigured the entire hill to build the city, slicing out terraces and platforms. By leveling the entire summit, they created a fifty-five-acre terrace half the size of the Vatican. At its zenith, Monte Alban housed seventeen thousand people and was by a considerable margin the biggest and most powerful population center in Mesoamerica.

The rationale for its construction is the subject of yet another lengthy archaeological dispute. One side proposes that Monte Alban formed because maize agriculture allowed the Oaxaca Valley's population to grow so much that the rural villages naturally clustered into something resembling cities. For most of its history Monte Alban was thus a huge village, not a true city, and certainly not a hierarchical state. Others argue that warfare had grown so devastating, as shown by the destruction of San Jose Mogote, that the main valley chiefdoms formed a defensive confederation headquartered at Monte Alban. Yet a third theory is that the Zapotec of Monte Alban-not the Olmec of La Venta-consolidated to form North America's first imperialist power, an aggressive state that subjugated dozens of other villages.

Among the strongest evidence for the last view are the nearly three hundred carved stone slabs at Monte Alban that depict slain, mutilated enemies: the rulers, Marcus believes, of communities conquered by Monte Alban. Some of the stones are labeled with enemy names, as with the unfortunate 1-Earthquake. These may commemorate victories in Monte Alban's grinding battle for supremacy with its local rival, San Martin Tilcajete, in the southern arm of the Central Valley. When San Jose Mogote founded Monte Alban, Tilcajete responded by gathering people from its surrounding villages, doubling in size, and erecting its own ceremonial buildings. War was the inevitable result. Monte Alban sacked Tilcajete in about 375 B.C. Undiscouraged, Tilcajete rebuilt itself on a better defensive position and acquired larger armies. When it again became a threat, Monte Alban attacked for the second time in 120 B.C. This time its forces finished the job. They burned the king's palace to the ground and emptied the rest of Tilcajete, leaving Monte Alban firmly in control of the entire valley.

With nothing to impede it, Monte Alban swept out and established a domain of almost ten thousand square miles. For centuries it stood on equal ground with its neighbors, the rising Maya states to the east and Teotihuacan to the north. It enjoyed relatively peaceful relations with both but had continual trouble with the nudzahui (pronounced "nu-sa-wi"-Spaniards called them the Mixtec), a constellation of petty principalities immediately to the west. By contrast with Monte Alban, these were minuscule entities; most were clusters of rustic villages covering ten to twenty square miles. Yet they were amazingly troublesome. Monte Alban repeatedly overran the nudzahui statelets, but never managed to eliminate them. These tiny, fractious domains endured for more than a thousand years. Meanwhile the much stronger and more centralized Zapotec empire collapsed completely in about 800 A.D.

nudzahui writing survives in eight codices, the deerskin or bark "books" whose painted pages could be folded like screens or hung on the wall like a mural. (The Spaniards destroyed all the rest.) More purely pictorial than Zapotec or Maya script, the texts were arranged almost randomly on the page; red lines directed the reader's eye from image to image. The symbols included drawings of events, portraits labeled by name (the king 4-Wind, for example, being shown by symbolic wind and four little bubbles in a line), and even punning rebuses. Enough writing has survived to give, when coupled with archaeological studies, a vivid picture of nudzahui life.

Like medieval Italian city-states, nudzahui principalities were rigidly stratified, with the king and a small group of kinspeople and noble advisers gobbling up much of the wealth and land. They constantly shifted configuration, some expanding by swarming over their neighbors, others imploding when their constituent villages seceded and joined other polities. More commonly, two states joined when their rulers married. Alliance through royal marriage was as common in eleventh-century Mixteca as it was in seventeenth-century Europe. In both, royal family trees formed an intricate network across national boundaries, but in Mixteca the queen's lands stayed in her line-the king's heir wasn't necessarily the queen's heir. Another difference: primogeniture was not expected. If the queen did not think her eldest son was fit for the crown, she could pass it to another child, or even to a nephew or cousin.

No fewer than four of the codices treat the story of 8-Deer Jaguar Claw, a wily priest-general-politician with a tragic love for the wife of his greatest enemy. Born in 1063 A.D., 8-Deer was a shirttail cousin to the ruling family of Tilantongo, which had been engaged for decades in a dynastic struggle with the kingdom of Red and White Bundle. (The name, a modern invention, comes from its name-glyph, which pictures the cloth wrapping used by the nudzahui to wrap holy objects; its exact location is still not nailed down.) Like his father, a high cleric, 8-Deer was trained for the priesthood, but political events and his own overweening ambition stopped him from following that path.

After an unprovoked attack on Red and White Bundle by Tilantongo raised hostilities to a fever, the warring parties agreed to meet in a sacred mountain cave with the Priestess of the Dead, a powerful oracle who had stripped away the flesh from her jaw, giving her a terrifying, skull-like appearance. Tilantongo's representative was 8-Deer, who attended the meeting in place of his recently deceased father. To his dismay, the priestess sided with Tilantongo's enemies and ordered 8-Deer, Tilantongo's champion, to exile himself a hundred miles away, in a jerkwater town on the Pacific called Tututepec.

Tucked away in Tututepec, 8-Deer assembled a private army, staffed it with many relatives, and in a series of swift campaigns seized dozens of neighboring villages and city-states. In addition to assembling the greatest empire ever seen in the region, the conquests managed to kill off most of the siblings and cousins above him in the line of royal succession. After six years of war he returned home to Tilantongo. During this visit, according to John M. D. Pohl, the archaeologist whose interpretations I am mostly following here, 8-Deer accidentally encountered 6-Monkey, the young wife of the much older king of Red and White Bundle. Despite the long enmity between the two kingdoms, 8-Deer and 6-Monkey secretly became lovers.

In 1096 Tilantongo's sovereign died in mysterious circumstances. The Priestess of the Dead selected 8-Deer's beloved elder half brother to be the regent-that is, the half brother became the last person between 8-Deer and the throne of Tilantongo. Three years later, unknown assailants stabbed the half brother to death in a sweatbath. The inconsolable 8-Deer took the throne of Tilantongo and declared war on Red and White Bundle, which he claimed had orchestrated the murder.

In this fragment from a nudzahui codex, the jaguar-cowled Lord 8-Deer (right) captures 4-Wind, son of his former lover, by the hair. As in other nudzahui codices, the characters' names are indicated by the accompanying circles-plus-head symbols.

Red and White Bundle's royal palace was built on a cliff over a bend in the river. Guarded by sheer walls on three sides, its soldiers had only to watch the fourth side, across which was an earthen berm. Leading an army of a thousand, 8-Deer threw up ladders, swarmed over the berm with his men, and entered the palace. As befit a conqueror, 8-Deer was wearing elaborate cotton armor, a ceremonial beard wig, and a cowl made from the head of a jaguar. Gold-and-jade necklaces dangled across his naked chest. In the palace he found 6-Monkey and her husband, the king of Red and White Bundle. Both were mortally wounded. In Pohl's account, 8-Deer held 6-Monkey as she died.