When Moses and company arrived, Canaan was a collection of citystates subjugated by Egyptian military superiority. Heavily fortified Canaanite cities controlled the agricultural lowlands. Undeterred, the new arrivals farmed the vacant uplands. "But the mountain-country shall be yours. Although it is wooded, you shall cut it down, and its farthest extent shall be yours" (Joshua 17:18). Settling in small villages, they cleared the forests and farmed terraced slopes in the hill country to gain a foothold in the Promised Land.

The Israelites adopted traditional Canaanite agriculture in their new hillside farms, growing what their neighbors grew. But they also practiced crop rotation and fallowing and designed systems to collect and deliver rainwater to their terraced fields. With the development of new iron tools, greater harvests led to agricultural surpluses that could support larger settlements. Leaving fields fallow every seventh year was mandated and animal dung was mixed with straw to produce Compost. 12 Land was regarded as G.o.d's property entrusted to the people of Israel for safekeeping. In the Judean highlands, Lowdermilk noted how a few well-maintained stone terraces still held soil after several thousand years of cultivation.

Agriculture expanded so much under the later Roman occupation that the empire's Middle East provinces were completely deforested by the first century AD. Grazing typically replaced forests on terrain too steep to farm. Throughout the region, flocks of goats and sheep reduced vegetation to stubble. Catastrophic soil erosion followed when too many livestock grazed steep hillsides. Forest soils built up over millennia disappeared. Once the soil was gone, so was the forest.

In a radio address from Jerusalem in June 1939, Walter Lowdermilk offered an eleventh commandment he imagined Moses might have slipped in had he foreseen what was to become of this promised land. "Thou shalt inherit the Holy Earth as a faithful steward, conserving its resources and productivity from generation to generation. Thou shalt safeguard thy fields from soil erosion ... and protect thy hills from overgrazing by thy herds, that thy descendants may have abundance forever. If any shall fail in this stewardship of the land ... thy descendants shall decrease and live in poverty or perish from off the face of the Earth."13 As Marsh had before him, Lowdermilk worried about the implications of what he saw in the Middle East for America's long-term prospects. Both looked to the Old World for lessons for the New World. Neither realized that the scenarios they worried about had already happened in America. Mayan civilization provides the best-studied, but by no means the only, example of soil degradation contributing to the collapse of a society in the Americas. The earliest Mayan settlements grew from the lowland jungle of the Yucatan Peninsula and slowly consolidated into increasingly complex settlements. By the second century BC, large ceremonial and commercial centers like Tikal coalesced into a complex hierarchical society of citystates with a common language, culture, and architecture. Mayan cities were comparable in size to Sumerian city-states. At its peak Tikal was home to from thirty thousand to fifty thousand people.

The first settled Mesoamerican communities grew to regional prominence after maize was domesticated about 2000 BC. Over the next thousand years, small villages came to depend on cultivating maize to supplement hunting and gathering from the wild lands between villages. Small-scale forest clearing for agriculture gradually expanded and maize became an increasingly important component of the Mesoamerican diet. As in Mesopotamia, dispersed networks of settlements grew into ceremonial centers and towns with priests, artisans, and administrators to oversee the redistribution of surplus food.

Initially, the productivity of domesticated maize was similar to wild varieties, which could be readily gathered. Tiny cobs about the size of a human thumb were simply chewed. People began grinding maize into flour once high-yield varieties allowed the development of permanent settlements. The region supported a diffuse rural population until large towns emerged between 350 Bc and AD 250. By then some parts of the Mayan world were already severely eroded, but in many areas the greatest soil erosion-and evidence for soil conservation efforts-date from about AD 600-900. The subsequent lack of artifacts has been interpreted to show a dramatic population decline (or dispersal) as Mayan society crumbled and the jungle reclaimed Tikal and its rivals.

Mayan population grew from less than two hundred thousand in 6oo BC to more than a million by AD 300. Five hundred years later at the peak of Mayan civilization, the population reached at least three million and perhaps as many as six million. Over the next two hundred years the population fell to less than half a million people. When John Stephens rediscovered ruined Mayan cities the region appeared deserted except for areas on the edge of the jungle. Even today the population density in the rapidly growing region remains below that of ancient times. So what happened?

Mayan agriculture began with a system known as slash-and-burn agriculture, in which a patch of jungle was cleared with stone axes and then burned before the onset of the rains, when maize and beans were planted. Ash from burning the cleared forest fertilized the soil and guaranteed good crops for a few years, after which fertility of the nutrient-poor tropical soil fell rapidly. Cleared patches could not be farmed for long before being abandoned to the jungle to restore soil fertility. A lot of jungle was needed to keep a few fields under cultivation. As in ancient Greece and Italy, the first evidence for extensive soil erosion coincides with pioneer farming.

Slash and burn worked well while the population density remained low and there was enough land for farmers to move their fields every few years. As the great Mayan cities rose from the jungle, people kept clearing land as their ancestors had done, but they stopped moving their fields. The tropical soils of the Yucatan Peninsula are thin and easily eroded. Under sustained cultivation, the high productivity obtained right after clearing and burning rapidly declines. Compounding this problem, the lack of domesticated animals meant no manure for replenishing the soil. Just as in Greece and Rome, rising demand for food and declining productivity compelled cultivation of increasingly marginal land.

After about 300 BC the region's increasing population led people to begin farming poorly drained valley bottoms and limestone slopes with thin, fragile soils. They built raised fields in swamps by digging networks of drainage ca.n.a.ls and piling up the excavated material in between to create raised planting beds perched above the water table. In some areas extensive terracing began around AD 250 and then spread across the landscape as the population continued to expand for another six and a half centuries. Mayan farmers terraced hillsides to create flat planting surfaces, slow erosion by overland flow, and divert water to fields. However, in major areas like Tikal and Copan there is little evidence for soil conservation efforts. Even with erosion control efforts, deposition of soil eroded from surrounding slopes disrupted wetland agriculture practiced in sinkholes.

Sediment cores from lakes in the Mayan heartland suggest that agricultural intensification increased soil erosion. The rate that sedimentation piled up on lakebeds increased substantially from 250 BC through the ninth century AD. While not necessarily responsible for the collapse of Mayan society, soil erosion peaked shortly before Mayan civilization unraveled about AD goo when the food surpluses that sustained the social hierarchy disappeared. Some Mayan cities were abandoned with buildings half finished.

In the 199os geographers studying small depressions, known as bajos, around Mayan sites in northwestern Belize found that cultivated wetlands had filled with soils eroded after deforestation of the surrounding slopes. The southern Yucatan is broken into depressions that formed natural wetlands extensively cultivated during the peak of Mayan civilization. Trenches revealed buried soils dating from the pre-Mayan period covered by two and a half to six feet of dirt eroded from the surrounding slopes in two distinct episodes. The first corresponded to forest clearing during the spread of pioneer farmers from the valleys up onto the surrounding hillslopes. The second occurred during agricultural intensification immediately before the end of Mayan civilization, after which soil began to rebuild as the forest reclaimed fields and wetlands.

Researchers also found evidence for accelerated soil erosion caused by extensive deforestation of sloping land in the Mayan lowlands. Where Mayan terraces remain intact, they hold three to four times more soil than lies on adjacent cultivated slopes. Development of erosion control methods allowed the Mayan heartland to support large populations but the expansion depended on intensive cultivation of erosion-p.r.o.ne slopes and sedimentation-p.r.o.ne wetlands. Eventually, Mayan civilization reached a point where its agricultural methods could no longer sustain its population.

Modern deforestation in the Peten is beginning to repeat the cycle of erosion after a thousand years of soil development. Since the early i98os landless peasant farmers have turned much of the region's forest into traditional Mayan milpas (small cultivated fields). A twentyfold increase in population from 1964 to 1997 has transformed the region from nearly unbroken forest to a nearly deforested landscape.

Soils on most of the region's hillslopes consisted of an organic horizon above a thin mineral soil sitting directly on weakly weathered limestone bedrock. One study found that under the region's last virgin forest, hillslope soils were about ten to twenty inches thick, whereas modern cultivated fields are already missing three to seven inches of topsoil-most of the 0 and A horizons. In some places, the rapid erosion following modern slope clearing and cultivation had already stripped the soil down to bedrock. Another study of soil erosion following modern forest clearance in central Belize found that one to four inches of topsoil were lost in a single ten-year cycle of clearing for corn and ca.s.sava for two to three years of cultivation followed by a fallow year. Complete removal of the soil would require just four milpa cycles. In a particularly striking example from the northern Yucatan, about eight inches of soil on the upland surfaces around a sinkhole named Aguada Carolina was eroded to bedrock in a decade of renewed cultivation. Similarly, researchers investigating ancient soil ero sion in the Mayan of the Peten noticed that the soil was stripped down to bedrock on newly cleared slopes in under a decade.

Rates of soil formation in the Central American jungle are far slower than rates of erosion under Mayan agriculture. The region's limestone bedrock weathers about half an inch to five inches in a thousand years. An average soil depth of about three inches developed on Mayan architecture abandoned a thousand years ago indicates rates of soil formation similar to the geologic erosion rate. Both are about a hundred times slower than erosion from cultivated slopes.

The Mayan heartland was not the only place where soil influenced Native American civilizations. Soils of central Mexico tell similar stories of severe erosion on steep hillslopes undermining agriculture.

In the late 1940s UC Berkeley professor Sherburne Cook drove around the central Mexican plateau and concluded that the land was in poorest condition in areas that had supported the largest populations before the Spanish conquest. The thick soil and sod covering uncultivated areas contrasted with the truncated soil profiles, slopes stripped down to weathered rock, and thick, artifact-rich valley fills derived from former hillslope soils that characterized more densely populated areas. Cook saw evidence for two periods of erosion, an ancient episode that stripped soils from hillsides and a more recent episode that entrenched deep gullies into the valley bottoms. "Evidently the entire range was once cleared for cultivation, abandoned, allowed to become covered with young forest, and finally, the lower portion again cleared."14 Despite Cook's revelation, the timing of these cycles remained uncertain until development of radiocarbon dating in the 1950s.

a.n.a.lyses of sediment cores from Lake Patzcuaro, east of Mexico City in Michoacan, revealed evidence for three distinct periods of rapid soil erosion. The first period accompanied extensive land clearance about 3,500 years ago shortly after maize cultivation began. The second period of high erosion occurred in the late precla.s.sic period between 2,500 and 1,200 years ago. The third erosional period peaked immediately before the Spanish conquest, when up to a hundred thousand people lived around the lake. Despite the introduction of the plow, soil erosion rates dropped as diseases decimated the region's population after Cortez arrived in AD 1521.

Just as in ancient Greece and around the Mediterranean, cycles of soil erosion in different parts of central Mexico occurred at different times and so were not driven by changes in climate. For example, in the PueblaTlaxcala area of the central Mexican highland, accelerated erosion of hill slope soils around 700 BC coincided with rapid expansion of settlements. A period of soil formation and cultural stagnation beginning about AD 100 was followed by a second period of accelerated erosion and rapid expansion of settlements before the Spanish conquest. Geoarchaeological surveys of hillslopes in the region revealed that agricultural fields were abandoned owing to soil erosion progressively from top to bottom, much as in ancient Greece.

As in the bajos of the Mayan jungle, pits dug into the swampy valley bottom sediments in the Upper Lerma Basin in central Mexico also record increased soil erosion from the surrounding slopes beginning around 1100 BC. Soil erosion then intensified during expansion of settlements in the late cla.s.sic and early postcla.s.sic periods beginning about AD 6oo. On his tour fourteen centuries later, Cook recognized that the areas populated most densely in preconquest times had the worst soil exhaustion.

The soils in Mexico's "cradle of maize" in the Tehuacan Valley about a hundred miles southeast of Mexico City also bear witness to extensive preColumbian soil erosion. In the early 199os field surveys of soils around the town of Metzontla revealed striking differences between areas that had been cultivated and those that had not. Intensively cultivated hillslopes were extensively eroded, with thin soils above weathered rock. Remnants of subsoil exposed at the ground surface doc.u.mented soil erosion from fields, leaving a modern soil that consists of little more than a thin mantle of broken rock. In contrast, areas with little evidence of past cultivation contained a foot and a half of well-developed soil above weathered rock. Abrupt transitions in soil depth between long-cultivated and uncultivated areas suggest that a foot and a half of soil was missing from the farms.

Expansion of agriculture from irrigated valley bottoms up onto the surrounding hillslopes about 1,300 to 1,700 years ago supported a growing population and triggered widespread soil erosion that still impoverishes the region. Agriculture on the slopes of the region today supplies only about a quarter of the small town's maize and beans. Metzontla's residents produce handcrafted ceramics or work at wage labor in other towns. In this semiarid area where soil production proceeds slowly, the residents' primary environmental concern is access to firewood for domestic use and for firing ceramics. Their soil disappeared slowly enough that they don't know it is gone.

Erosion from agriculture also caused abandonment of parts of southern Central America. Pollen from a long core pulled from the bottom of La Yeguada's small lake in central Panama records that the rainforest was cleared for slash-and-burn agriculture between 7,000 and 4,000 years ago. Archaeological records from this period indicate considerable population growth as intensified agriculture stripped the forest from the lake's watershed. By the time of Christ, accelerated erosion in the foothills and uplands led to agricultural abandonment of the watershed. Slow forest regeneration suggests depleted soils, and later agricultural settlements were concentrated along previously unoccupied floodplains and coastal valleys. The uppermost layers in the long sediment core revealed that the primeval rainforest actually dates from the time of the Spanish conquest, when the indigenous population of the area again declined dramatically-this time felled by disease.

In the American Southwest, the spectacular ruins of Mesa Verde, Chaco Canyon, and Canyon de Ch.e.l.ly-all abandoned well before discovery by Euro-Americans-have long intrigued archaeologists. Between about AD 1250 and 1400, the native Pueblo culture vanished from the Southwest. The usual suspects of war, disease, drought, and deforestation have been proposed to explain the mystery.

Pollen sequences recovered from different depths in valley bottom sediments show little to no change in the vegetation community at Chaco Canyon for thousands of years-until the Pueblo people arrived. Plant remains preserved in crystallized packrat urine built up on the floor of caves show that the native vegetation was pinyon-juniper woodland, and that the local vegetation changed dramatically during Pueblo occupation. The inhabitants of Chaco Canyon used thousands of ponderosa pines to construct buildings between AD 1000 and 1200. Countless more trees were burned as fuel. Today the local vegetation on most of the valley floor is a mix of desert scrub and gra.s.ses. But if you hike near the canyon you can still see ancient stumps in areas where few trees now grow.

Many have argued that droughts led to the abandonment of Chaco Canyon. Although droughts probably contributed to the Pueblo culture's decline, the regional climate for the past thousand years falls within the range of variability for the past six thousand years. It seems more likely that salinization of the Pueblos' fields and soil erosion limited the life span of their agriculture as a growing population led to dependence on neighboring areas for basic resources. These conditions set up an agricultural disaster during the next drought.

Domesticated maize arrived at Chaco Canyon about 1500 BC. Initially grown near ephemeral streams or freshwater marshes, maize production increasingly depended on floodplain irrigation as agriculture expanded. By about AD 800 to 1000, rain-fed farming was practiced wherever feasible throughout the Southwest. Agricultural settlements ranged in size from small communities of a few dozen people to villages with hundreds of inhabitants. Foraging remained an important component of the dietparticularly during droughts.

At first sites were occupied for a few decades before people moved on to new locations, but by about AD 1150 there was no unused arable land to move into or cultivate when local crops failed. The landscape was full, the desert's rainfall was capricious, and its soil was fragile. As in the Old World centuries earlier, settlements became increasingly sedentary and their heavy investment in agricultural infrastructure discouraged farmers from leaving fields fallow every few years. Beginning about AD 1130, two centuries of drought and chaotic rainfall patterns occurred while all the arable land was already under cultivation. When crop failures on marginal land forced people to move back into more settled areas, the remaining productive land could not support them.

Comparison of ancient agricultural soils and uncultivated soils in New Mexico and Peru shows that agricultural practices need not undermine societies. Soils at a site in the Gila National Forest, typical of prehistoric agricultural sites in the American Southwest, were cultivated between AD 1100 and 1150, at the peak of Pueblo culture, and subsequently abandoned. Soils of sites cultivated by the Pueblo culture are lighter colored, with a third to a half of the carbon, nitrogen, and phosphorus content of neighboring uncultivated soils. In addition, cultivated plots had gullies-some more than three feet deep-that began during cultivation. Even today, little gra.s.s grows on the ancient farm plots. Native vegetation cannot recolonize the degraded soil even eight centuries after cultivation ceased.

In contrast, modern farmers in Peru's Colca Valley still use ancient terraces cultivated for more than fifteen centuries. Like their ancestors, they maintain soil fertility through intercropping, crop rotations that include legumes, fallowing, and the use of both manure and ash to maintain soil fertility. They have an extensive homegrown system of soil cla.s.sification and do not till the soil before planting; instead they insert seeds into the ground using a chisel-like device that minimally disturbs the soil. These long-cultivated soils have A horizons that are typically one to four feet thicker than those of neighboring uncultivated soils. The cultivated Peruvian soils are full of earthworms and have higher concentrations of carbon, nitrogen, and phosphorus than native soils. In contrast to the New Mexican example, under traditional soil management these Peruvian soils have fed people for more than fifteen hundred years.

The contrast between how the Pueblos and the Incas treated their dirt is but another chapter in the broader story of how the rise of agriculture set off a perpetual race to figure out how to feed growing populations by continuously increasing crop yields. Sometimes cultures figured out a way to muddle through without depleting soil productivity, often they did not.

A common lesson of the ancient empires of the Old and New Worlds is that even innovative adaptations cannot make up for a lack of fertile soil to sustain increased productivity. As long as people take care of their land, the land can sustain them. Conversely, neglect of the basic health of the soil accelerated the downfall of civilization after civilization even as the harsh consequences of erosion and soil exhaustion helped push Western society from Mesopotamia to Greece, Rome, and beyond.

Efforts to feed the world today often include calls for a cultural revolution, a new agrotechnological revolution, or a political revolution to redistribute land to subsistence farmers. Less widely known is how, after centuries of agricultural decline, a preindustrial agricultural revolution began in the still-fertile and revitalized fields of Western Europe, setting the stage for the social, cultural, and political forces that forged colonial powers and shaped our modern global society.

LET THEM EAT COLONIES

There is nothing new except what has been forgotten.

MARIE ANTOINETTE.

GUATEMALANS GROW SOME OF THE BEST COFFEE in the world, but most can't buy it at home. Neither can tourists. When I was there last I had to wake up on freeze-dried Mexican Nescafe, even though I can buy bags of freshly roasted Guatemalan coffee beans two blocks from my house in Seattle. Less well known than the story of how Europe carved out global empires is how the way Europeans treated their soil helped launch the exploration and history of the New World. Today's globalized agriculture that ships local produce overseas to wealthier markets reflects the legacy of colonial plantations established to help feed European cities.

Like many ancient agricultural societies, Europeans began working to improve their dirt once soil fertility and access to fresh land declined. But unlike the Mediterranean's intense spring and summer rains that promoted erosion from bare fields, western Europe's gentle summer rains and winter-spring snow pack limited erosion of even highly erodible loess soils when farmed. Moreover, by rediscovering soil husbandry western Europeans kept soil degradation and erosion at bay long enough to establish colonial empires that provided new land to exploit.

Farming spread from the Middle East into Greece and the Balkans between seven and eight thousand years ago. After moving into central Europe's easily worked loess, agriculture steadily advanced north and west, reaching Scandinavia about three thousand years ago. Consuming Europe's forest soils as it went, agriculture left a record of boom-and-bust cycles a.s.sociated first with Neolithic and Bronze Age cultures, then Iron Age and Roman society, and most recently the medieval and modern periods when colonial empires began mining soil and sending both produce and profits back to feed Europe's increasingly urban populace-the Industrial Revolution's new cla.s.s of landless peasants.

The first agricultural communities reached Europe's doorstep in southern Bulgaria around 5300 BC. At first farmers grew wheat and barley in small fields surrounding a few timber-framed buildings. Agricultural expansion into marginal land lasted about two thousand years before the agricultural potential of the region was fully exploited and persistent cultivation began to exhaust the soil. With no evidence of a climate shift, local populations grew and then declined as agricultural settlement swept through the area. Evidence for extensive late Neolithic soil erosion shows that agriculture spread from small areas of arable soils on the valley bottoms into highly erodible forest soils on steeper slopes. Eventually, the landscape filled in with small communities of several hundred people farming the area within about a mile of their village.

In these first European communities, population rose slowly before a rapid decline that emptied settlements out for five hundred to a thousand years, until the first traces of Bronze Age cultures then appeared. This pattern suggests a fundamental model of agricultural development in which prosperity increases the capacity of the land to support people, allowing the population to expand to use the available land. Then, having eroded soils from marginal land, the population contracts rapidly before soil rebuilds in a period of low population density.

This roller-coaster cycle characterizes the relation between population and food production in many cultures and contexts because the agricultural potential of the land is not a constant-both technology and the state of the soil influence food production. Improved agricultural practices can support more people with fewer farmers, but soil health eventually determines how many people the land can support. Floodplains continually receive nutrients from periodic flooding, but most other land cannot produce continuously high crop yields without intensive fertilization. So once a society comes to depend on upland farming it can cultivate a fraction of its land base at any one time, expand the area under cultivation, keep inventing new methods to counteract declining soil fertility, or face agricultural decline owing to degradation of soil fertility or gradual loss of the soil itself.

As agriculture spread north and west, people opened the first clearings in Europe's ancient forest to cultivate small plots for a few years at a time. Ash from burned vegetation fertilized newly cleared fields, helping to maintain initial crop yields until soil fertility declined enough to make it worth the ha.s.sle of moving on. The practice of abandoning worn-out fields periodically left fallow land to revegetate, first with gra.s.ses, then shrubs, and eventually back to forest. Ground cultivated for a few years then lay fallow for decades as the recolonizing forest gradually revived the soil, allowing clearing and planting again decades later.

Lake sediments, floodplain deposits, and soils record the postglacial evolution of the European landscape. From 7000 to 5500 BC stable environmental conditions left little evidence of human impact. Pollen preserved in lakebeds shows that Neolithic farmers opened clearings in dense forest as agriculture spread north from the Balkans. Cereal pollen shows up in soil profiles and sediment cores about 5500 BC in central Europe. Sediment cores from lakes provide the first incontrovertible evidence of substantial human impact on the central European landscape as ma.s.sive amounts of charcoal and increased sedimentation-evidence for accelerated soil erosion-coincide with pollen evidence of extensive forest clearing and cereal cultivation around 4300 BC, when postglacial temperatures in Europe were at a maximum.

Farmers had arrived, but Europe was still wild. Lions and hippopotamuses lived along the Thames and Rhine rivers. While scattered bands of people foraged around Europe's lakes, rivers, and coast, a rich soil developed beneath huge oak, elm, and beech trees that stabilized loess-mantled slopes.

Germany's first farmers were drawn to the forest soil developed on silt dropped by glacial winds between the Rhine and Danube rivers. Several centuries later a second wave of related arrivals settled across northern Europe in a band stretching from Russia to France. Soon farmers grew wheat, barley, peas, and lentils on the region's fertile loess. Hunting and gathering thrived outside the loess belt.

Neolithic farmers kept livestock and lived in large longhouses near fields along rivers and streams. Houses were occupied for the several decades that the surrounding fields were kept under continuous cultivation. As isolated longhouses began coalescing into small hamlets, farming spread beyond the loess. More land was cleared and cultivated more continuously. By about 3400 Bc hunting for survival was history throughout central Europe.

German soils record periods of agriculturally induced soil erosion from hillsides followed by periods of soil formation lasting roughly five hundred to a thousand years. Soil profiles and alluvial sediments in southern Germany's Black Forest record several periods of rapid erosion a.s.sociated with increased population. Neolithic artifacts in truncated soil profiles show that initial erosion after the arrival of agriculture about 4000 BC culminated in extensive soil loss by 2000 BC. Declining cereal pollen and a period of soil formation characterized a thousand years of lower population density until renewed erosion in Roman times peaked in the first centuries AD. A second cycle of agricultural decline, soil formation, and forest expansion followed until renewed population growth in the Middle Ages initiated a third, ongoing cycle.

The soils at Frauenberg, a Neolithic site in southeastern Germany, record erosion of nearly the entire soil profile that began with early Bronze Age agriculture. Located on a hill that rises three hundred feet inside a bend in the Danube River, the site's combination of loess soils and a sweeping view of the surrounding country appealed to prehistoric farmers. Remnants of the original soil found in excavations at the site doc.u.ment three distinct periods of occupation corresponding to Bronze Age farming, a Roman fort, and a medieval monastery. Radiocarbon dating of charcoal pulled from soil horizons show that little erosion occurred as soil developed after deglaciation-until Bronze Age farming exposed clay-rich subsoil at the ground surface and eroded nearly the entire loess cover. Subsequent erosion slowed once the less erodible subsoil was exposed. Forest currently blankets the site, which still has limited agricultural potential.

Evidence from soils, floodplains, and lake sediments at sites across Germany shows that human impact has been the dominant influence on the landscape since the last glaciation. Erosion and human occupation occurred in tandem but not in regional pattern as expected for climate-driven events. Just as in ancient Greece and around the Mediterranean, central European cycles of agricultural clearing and erosion a.s.sociated with population growth gave way to migration, population decline, and renewed soil formation.

Surveys of truncated hillslope soils at more than eight hundred sites along the Rhine River indicate that post-Roman agriculture stripped up to several feet of soil from hillslopes cleared of native forest. Erosion since AD 6oo has been about ten times the erosion rate before forest clearing through erosive runoff across bare, plowed fields. Similar soil surveys in Luxembourg report an average of twenty-two inches of lost soil and accelerated soil loss over more than 9o percent of the landscape. Despite the prevalence of Neolithic farming on central Europe's slopes, most of the region's modern agricultural land lies on valley-bottom deposits of reworked soil eroded off surrounding slopes.

Neolithic settlements in southern France are concentrated almost exclusively on limestone plateaus known today for bare white slopes sporting thin, rocky soil and spa.r.s.e vegetation. When farmers arrived, these uplands were covered by thick brown soil that was far easier to plow than clay-rich valley bottoms. No longer suitable for cultivation, and considered something of a backwater, the limestone plateaus around Montpellier are used primarily for grazing. The harbor at nearby Ma.r.s.eille began filling with sediment soon after Greek colonists founded the city in 6oo BC. Sedimentation in the harbor increased thirtyfold after agricultural clearing spread up the steep slopes around the new town.

Early forest clearing in Britain led to extensive soil erosion long before the Roman invasion, as a growing population slowly cleared the forest to plow the slopes. High population density in Roman times exacerbated the loss, in part because better plows worked more of the landscape more often. The population fell dramatically as the empire collapsed, and took almost a thousand years to build back to the same level.

Floodplain sediments along Ripple Brook, a small tributary of the River Severn typical of lowland Britain, record a dramatic increase in deposition rates (and therefore hillslope soil erosion) in the late Bronze Age and early Iron Age. The relative abundance of tree pollen recovered from valley bottom sediments shows that between 2,900 and 2,500 years ago the heavily forested landscape was cleared and intensively farmed. A fivefold increase in floodplain sedimentation speaks to a dramatic increase in hillslope erosion.

Net soil loss averages between three and six inches since woodland clearance in England and Wales. Some watersheds have lost up to eight inches of topsoil. Although much of the loss occurred in the Bronze Age or Roman times, in some places substantial erosion occurred after medieval times. Just two hundred years after Nottinghamshire's famed Sherwood Forest was cleared for agriculture, the original forest soil has been reduced to a layer of thin brown sand over rock. Just as in Lebanon's ancient cedar forest, most of the topsoil is now gone from Robin Hood's woods.

Across the border in Scotland, radiocarbon dating of a sediment core recovered from a small lake west of Aberdeen provides a continuous record of erosion from the surrounding slopes for the past ten thousand years. Sediment deposition rates in the lake, and thus erosion rates on the surrounding slopes, were low for five thousand years under postglacial shrub land and birch forest. Following the arrival of agriculture, pollen from crops and weeds coincide with a threefold increase in the sediment deposition rate. After the Bronze and Iron Ages, erosion decreased dramatically for almost two thousand years as native plants regenerated across a largely abandoned landscape-until erosion accelerated again in the modern era.

Similar cores taken from small lakes in southern Sweden also record the transition from little preagricultural erosion to much higher rates after arrival of the plow. One from Lake Bussjosjo shows that forest stabilized the landscape from 7250 to 750 BC until erosion accelerated following forest clearing. Erosion increased further under intensified agriculture in the sixteenth and seventeenth centuries. A core pulled from Havgardssjon provides a 5,ooo-year record of vegetation and erosion. The archaeological record around the lake has no Bronze or Iron Age artifacts. Lakebed sediments piled up four to ten times faster after agricultural settlement began around AD iioo. All across the glaciated terrain of Scandinavia, Scotland, and Ireland, farmers could not make a living until enough ice-free time pa.s.sed to build soil capable of sustaining cultivation.

Put simply, European prehistory involved the gradual migration of agricultural peoples, followed by accelerated soil erosion, and a subsequent period of low population density before either Roman or modern times. Just as in Greece and Rome, the story of central and western Europe is one of early clearing and farming that caused major erosion before the population declined, and eventually rebounded.

As the Roman Empire crumbled, the center of its civilization shifted north. Abandoning Rome as the capital, Diocletian moved his government to Milan in AD 300. When Theodoric established the Gothic kingdom of Italy on the ruins of the Roman Empire, he chose Verona as his new capital in the north. Even so, many of northern Italy's fields lay fallow for centuries until an eleventh-century program of land reclamation began returning them to cultivation. After several centuries of sustained effort, most of northern Italy's arable land was again under cultivation, supporting prosperous medieval cities that nurtured a renaissance of literature and art.

As northern Italy's population rebounded, intensive land use increased silt loads in the region's rivers enough to attract the attention of Leonardo da Vinci and revive the Roman art of river engineering and flood control. Intensive cultivation on hillside farms spread into the Alps, producing similar results on the Po River as Roman land use had on the Tiber River. Eventually, after eight centuries of renewed cultivation, even northern Italy's soil faltered. Mussolini's Fascist government spent about half a billion dollars on soil conservation in the 1930s.

Because Rome imported most of its grain from North Africa, Egypt, and the Middle East, it made fewer demands on the soils of the Po Valley, Gaul (France), Britain, and the Germanic provinces. Roman agriculture in its western European provinces was mostly confined to river valleys; for the most part hillslopes that had been farmed in the Bronze Age remained forested until medieval times. It is no coincidence that these northern provinces fed the western European civilization that centuries later rose from the ruins of the Roman Empire.

After the empire collapsed, many Roman fields north and west of the Alps reverted to forest or gra.s.s. In the eleventh century, farmers worked less than a fifth of England. With half in pasture and half in crops left fallow every other year, this meant that only about 5 percent of the land was plowed each year. Less than io percent of Germany, Holland, and Belgium were plowed annually in the Middle Ages. Even in the most densely populated parts of southern France, no more than 15 percent of the land was cultivated each year.

In early medieval times, townships controlled a given area of land held in common by all villagers. Each household received a share of land to cultivate each season, after which the fields reverted to communal use. The general rule was to plant a crop of wheat, followed by beans and then a fallow season. After the harvest, cattle wandered the fields turning crop stubble into meat, milk, and manure.

Columbia University professor Vladimir Simkhovitch saw the structure of medieval village communities as an adaptation to farming degraded soils. He noted that a similar pattern of land use and ownership characterized many old villages throughout Europe where the land holdings of individual peasants had not been fenced off and enclosed. Barns, stables, and vegetable gardens were always near homesteads, but fields were divided into a patchwork of land belonging to individual farmers. Each farmer might own ten or more parts of three different fields managed collectively for a crop of wheat or rye, then oats, barley, or beans, and finally fallow pasture.

Simkhovitch argued that an inconvenient arrangement in which a farmer had no say in the rotation or type of tillage used on his fieldswhich could be quite distant from each other-must have been adopted throughout the continent for good reason. He doubted that such arrangements were simply inherited from Roman villas or imposed under feudal ism. Simkhovitch hypothesized that an individual farmer could not keep enough cattle to maintain the fertility of his plot, but a village's livestock could collectively fertilize the commons enough to slow their degradation. Simkhovitch believed that the already degraded state of the land made cooperation the way to survive-a notion contrary to the "tragedy of the commons" in which collective farming was thought to have caused land degradation in the first place.

Figure to. Miniature from an early-sixteenth-century ma.n.u.script of the Middle English poem G.o.d Spede ye Plough (original held at the British Museum).

Simkhovitch argued that by failing to maintain their soil, ancient societies failed themselves. "Go to the ruins of ancient and rich civilizations in Asia Minor, Northern Africa or elsewhere. Look at the unpeopled valleys, at the dead and buried cities.... It is but the story of an abandoned farm on a gigantic scale. Depleted of humus by constant cropping, land could no longer reward labor and support life; so the people abandoned it."' The introduction of alfalfa and clover into European agriculture helped rebuild soil fertility, Simkhovitch insisted. Noting that there were no hay fields before the sixteenth and seventeenth centuries, he suggested that enclosure of common fields allowed converting enough land to pasture to raise the cattle and sheep needed to manure the land and thereby increase crop yields.

The conventional explanation for the low crop yields of medieval agriculture invoked a lack of enough pasture to supply cultivated land with the manure needed to sustain soil fertility. Until recently, historians generally considered this to reflect ignorance of the value of manure in maintaining soil fertility. It now seems as likely that medieval farmers knew that keep ing land in pasture restored soil fertility, but impatience and economics made the required investments unattractive to folks perpetually focused on maximizing this year's harvest.

After centuries when post-Roman agricultural methods and practices limited crop yields, population growth accelerated when an extended run of good weather increased crop yields during medieval times. As the population grew, the clearing of Europe's remaining forests began again in earnest as new heavy plows allowed farmers to work root-clogged lowlands and dense river valley clays. From the eleventh to the thirteenth century, the amount of cultivated land more than doubled throughout western Europe. Agricultural expansion fueled the growth of towns and cities that gradually replaced feudal estates and monasteries as the cornerstone of Western civilization. Europe's best soils had been cleared of forest by about AD 12oo. By the close of the thirteenth century, new settlements began plowing marginal lands with poor soils and steep terrain. Expansion of the area of planted fields allowed the population to keep growing. Doubling over a couple of centuries, by AD 1300 Europe's population reached eighty million.

Powerful city-states arose where the most land was plowed under, particularly in and near the fertile lowlands of Belgium and Holland. By the middle of the fourteenth century, farmers were plowing most of western Europe's loess to feed burgeoning societies and their new middle cla.s.s. Already hemmed in by powerful neighbors, Flemish and Dutch farmers adopted crop rotations similar to those still used today.

The catastrophic European famine of 1315-17 provides a dramatic example of the effect of bad weather on a population near the limit of what its agricultural system could support. Every season of 1315 was wet. Waterlogged fields ruined the spring sowing. Crop yields were half of normal and what little hay grew was harvested wet and rotted in barns. Widespread food shortages in early 1316 compelled people to eat the next year's seed crop. When wet weather continued through the summer, the crops failed again, and wheat prices tripled. The poor could not afford food and those with money-even kings-could not always find it to buy. Bands of starving peasants turned to robbery. Some even reportedly resorted to cannibalism in famine-stricken areas.

Malnutrition and starvation began to haunt western Europe. The population of England and Wales had grown slowly but steadily after the Norman invasion until the Black Death of 1348. Major famines added to the toll. The population of England and Wales fell from about four million in the early 13006 to about two million by the early 1400s. Europe's population dropped by a quarter.

After the Black Death depopulated the countryside, landlords competed to retain tenant farmers by granting them lifelong or inheritable rights to the land they worked in exchange for modest rents. As the population rebounded, a final push of agricultural expansion filled out the landscape with farms in the early sixteenth century. Starting in the late 15005 landlords motivated by the promise of getting higher rents from leasing land at inflated rates began enclosing lands formerly grazed in common. Already out of land and surrounded by powerful neighbors, the Dutch started their ambitious campaign to take land from the sea.

John Fitzherbert's 1523 Book of Surveying, the first work on agriculture published in English, held that the way to increase the value of a township was to consolidate rights to common fields and pasture into single enclosed tracts next to each farmer's house. Over the next several centuries this idea of reorganizing the commons to give every farmer three acres and a cow evolved into transforming the English countryside into large estates, portions of which could be rented out profitably to tenant farmers. Except for the peasants working the land, most thought that privatizing the commons would injure none, and benefit all by increasing agricultural production.

In the tumultuous sixteenth and seventeenth centuries much of England's agricultural land changed hands in Henry VIII's war against the Catholic Church, the wars of succession, and the English Civil War. The insecurity of land tenure discouraged investing in land improvement. By the second half of the seventeenth century, some argued that England should adopt the Flemish custom of agricultural leases under which the owner would pay a specified sum to the tenant if four impartial persons, two selected by the landlord and two by the tenant, agreed that the soil had been improved at the end of the lease.

As Europe's climate slid from the medieval warm period into the Little Ice Age (which lasted from about AD 1430 to 1850), extended cold periods meant shorter growing seasons, reduced crop yields, and less arable land. Perennially living on the edge, the lower cla.s.ses were vulnerable to severe food shortages after bad harvests. Governments monitored the price of bread to gauge the potential for social instability.

Desire for land reform among the peasantry, fueled by instability and shortages, would help trigger the Reformation. Land held by the Church had grown over the centuries far beyond the fields cleared by monks, because the Church seldom relinquished land bequeathed by the faithful. Instead, bishops and abbots rented out G.o.d's land to poor, land-hungry peasants. By the fifteenth century the Church, which owned as much as four-fifths of the land in some areas, overtook the n.o.bility as Europe's largest landlord. Monarchs and their allies seeking to seize church lands harnessed widespread resentment among tenants. Popular support for the Reformation rested as much on desire for land as the promise of religious freedom.

An increasing demand for crops meant less pasture, little overwinter animal fodder, and not enough manure to sustain soil fertility. As the population kept rising, intensively cultivated land rapidly lost productive capacity-increasing the need to plant more marginal land. Shortage of vacant land to plow helped motivate the rediscovery of Roman agricultural practices such as crop rotations, manuring, and composting.

Renewed curiosity about the natural world also stimulated agricultural experimentation. In the sixteenth century, Bernard Palissy argued that plant ashes made good fertilizer because they consisted of material that the plants had pulled from the soil and could therefore reuse to fuel the growth of new plants. In the early 16oos Belgian philosopher Jan Baptista van Helmont tried to settle the question of whether plants were made of earth, air, fire, or water. He planted a seedling tree in two hundred pounds of soil, protected it from dust, and let it grow for five years, adding only water. Finding that the tree had grown by one hundred and seventy pounds, whereas the soil had lost an insignificant two ounces, van Helmont concluded that the tree had grown from water-the only thing added to the process. Given that the soil had lost but a minuscule fraction of the weight of the tree, he dismissed the potential for earth to have contributed to the tree's growth. I doubt that he ever seriously considered air as a major contributor to the ma.s.s of the tree. It took a few more centuries before people discovered carbon dioxide and came to understand photosynthesis.

In the meantime, agricultural "improvers" came to prominence in the seventeenth century once the landscape was fully cultivated. Most of the low hills and shallow valleys of the Netherlands are covered by quartz-rich sand ill suited for agriculture. Supporting a growing population on their naturally poor soils, the Dutch began mixing manure, leaves, and other organic waste into their dirt. Working relatively flat land where erosion was not a problem, over time they built up dark, organic-rich soils to as much as three feet thick. Lacking more land, they made soil. As the Dutch had, the Danes improved their sandy dirt enough to more than double their harvests by adopting crop rotations including legumes and manure. In other words, they readopted key elements of Roman agriculture.

Soil improvement theories spread to England where population growth motivated innovation to increase crop yields. Seventeenth-century agriculturalists broadened the range of fodder crops, developed more complex crop rotations, used legumes to improve soil fertility, and used more manure to maintain soil fertility. In addition, introduction of the Flemish practice of growing clover and turnips as ground cover and winter fodder changed the ratio of animals to land, increasing the availability of manure. Improvers promoted clover as a way to rejuvenate fields and regain high crop yields: clover increased soil nitrogen directly through the action of nitrogen-fixing bacteria in nodules on the plant roots and, as feed for cattle, also produced manure.

Despite cold winters, wet summers, and shorter growing seasons, English agriculture increased its yields per acre from 1550 to 1700, in what has been called the "yeoman's agricultural revolution." At the start of the seventeenth century, between a third and half of English agricultural land was held by yeomen-small freehold farmers and those with long-term leases. In the early 16oos farmers obsessed with fertilization began plowing into their fields lime, dung, and almost any other organic waste that could be obtained. Farmers also began shifting away from fixed grain lands and pasture and began planting fields for three or four years, and then putting them to pasture for four or five years before plowing them up once again. This new practice of "convertible husbandry" resulted in much higher crop yields, making it attractive to plow up pastures formerly held in common.

The new breed of land improvers also pioneered systems to drain and farm wetlands. They experimented with plow design and with ways to improve soil fertility. Upper-cla.s.s landowners advocated enclosing pasturelands and growing fodder crops (especially turnips) to provide winter feed for cattle and increase the supply of manure. Adopting the premise that communal land use degraded the land, an idea now called the "tragedy of the commons," agricultural improvers argued that enclosing the commons into large estates was necessary to increase agricultural output. A Parliament of property owners and lawyers pa.s.sed laws to fence off fields that had been worked in common for centuries. Land enclosures increased crop yields and created tremendous wealth for large landowners, but the peasants thus fenced out-whose parents ate meat, cheese, and vegetables raised by their own efforts-were reduced to a diet of bread and potatoes.

Soil husbandry began to be seen as the key to productive, profitable farming. Gervase Markham, one of the first agricultural writers to write in English instead of Latin, described soils as various mixtures of clay, sand, and gravel. What made good soil depended on the local climate, the character and condition of the soil, and the local plants (crops). "Simple Clays, Sands, or Gravels together; may be all good, and all fit to bring forth increase, or all ... barren." Understanding the soil was the key to understanding what would grow best, and essential to keeping a farm productive. "Thus having a true knowledge of the Nature and Condition of your ground.... it may not only be purged and clensed ... but also so much bettered and refined." 2 Prescribing steps to improve British farms, Markham recommended using the right type of plow for the ground. He advised mixing river sand and crushed burned limestone into the soil, to be followed by the best manure to be had-preferably ox, cow, or horse dung. In describing procedures for improving barren soils, Markham advocated growing wheat or rye for two years in a field, and then letting sheep graze and manure it for a year. After the sheep, several crops of barley were to be followed in the seventh year by peas or beans, and then several more years as pasture. After this cycle the ground would be much improved for growing grain. The key to sustaining soil fertility was to alternate livestock and crops on the same piece of ground.

Equally important, although it received less attention, was preventing erosion of the soil itself. Markham advised plowing carefully to avoid collecting water into erosive gullies. Good soil was the key to a good farm, and keeping soil on the farm required special effort even on England's gentle rolling hills.

Almost half a century later, on April 29, 1675, John Evelyn presented a "Discourse on Earth, Mould and Soil" to England's Royal Society for the Improvement of Natural Knowledge. In addressing what he feared could be considered a topic unworthy of the a.s.sembled luminaries he invited the society's fellows to descend from contemplating the origin of heavenly bodies and focus instead on the ground beneath their feet. He implored them to consider both how soil formed and how the nation's long-term prosperity depended on improving the kingdom's dirt.

Evelyn described how distinct layers of topsoil and subsoil developed from the underlying rock. "The most beneficial sort of Mould or Earth, appearing on the surface ... is the natural (as I beg leave to call it) underturf-Earth and the rest which commonly succeeds it, in Strata's or layers, 'till we arrive to the barren, and impenetrable Rock." Of the eight or nine basic types of soil, the best was the rich topsoil where mineral soil mixed with vegetation.

I begin with what commonly first presents it self under the removed Turf, and which, for having never been violated by the Spade, or received any foreign mixture, we will call the Virgin Earth; ... we find it lying about a foot deep, more or less, in our Fields, before you come to any manifest alteration of colour or perfection. This surface-Mold is the best, and sweetest, being enriched with all that the Air, Dews, Showers, and Celestial Influences can contribute to it.

The ideal topsoil was a rich mixture of mineral and organic matter introduced by "the perpetual and successive rotting of the Gra.s.s, Plants, Leaves, Branches, [and] Moss ... growing upon it."3 Regaling his audience with the works of Roman agriculturalists, Evelyn described how to improve soil with manure, cover crops, and crop rotations. Like the Romans, Evelyn used odor, taste (sweet or bitter), touch (slippery or gritty), and sight (color) to evaluate a soil. He described different types of manure and their effects on soil fertility, as well as the virtues of growing legumes to improve the soil.

Echoing Xenophon, Evelyn held that to know the soil was to know what to plant. One could read what would grow best by observing what grew naturally on a site. "Plants we know, are nourished by things of like affinity with the const.i.tution of the Soil which produces them; and therefore 'tis of singular importance, to be well read in the Alphabet of Earths and Composts." Because soil thickened as organic material supplied from above mixed with the rotting rocks below, sustaining good harvests required maintaining the organic-rich topsoil ideal for crops. Mineral subsoil was less productive, but Evelyn believed that nitrous salts could resuscitate even the most exhausted land. "I firmly believe, that were Salt Peter ... to be obtain'd in Plenty, we should need but little other Composts to meliorate our Ground."4 Well ahead of his time, Evelyn antic.i.p.ated the value of chemical fertilizers for propping up-and pumping up-agricultural production.

By the start of the eighteenth century, improving farmland was seen as possible only through enclosing under private ownership enough pasture to keep livestock capable of fertilizing the plowed fields. Simply letting the family cow p.o.o.p on the commons would not do. The need for manure imposed an inherent scale to productive farms. Too small a farm was a recipe for degrading soil fertility through continuous cropping. Although very large farms turned out to mine the soil itself, this was not yet apparent-and Roman experience in this regard was long forgotten. To the individual farmer, enclosure was seen as the way to ensure a return on investing to improve soil fertility from well-manured ground.

Figure ii. t.i.tle page to The Whole Art of Husbandry, published in 1708.

Agricultural writers maintained that the key to good crop yields was to keep an adequate supply of manure on hand-to keep the right ratio of pasture to field on each farm, or estate as the case increasingly became. "The Arable-land must be proportioned to the quant.i.ty of Dung that is raised in the Pasture, because proper Manure is the chief Advantage of Arable-ground."5 The key to increasing agricultural productivity was seen to lie in bringing stock raising and cereal production into proximity and returning manure to the fields.

Still, not all land was the same; improvements needed to be tailored to the nature of the soil. British farmland consisted of three basic types: uplands lying high enough not to flood, lowlands along rivers and wetlands, and land susceptible to inundation by the sea. These lands had different vulnerabilities.

On hillslopes, the thin layer of a foot or so of topsoil was essential to good farming. Such lands were naturally p.r.o.ne to erosion and vulnerable to poor farming practices. On lowlands, the soil was replenished by upland erosion that produced fine deposits downslope. "As to Lands lying near Rivers, the great Improvement of them is their over-flowing, which brings the Soil of the Uplands upon them, so as that they need no other mending though constantly mowed."6 Working land too hard for too long would reduce soil fertility. Sloping land was particularly vulnerable. "Where Lands lie upon the sides of Hills ... great care must be taken not to plow them out of heart." 7 Recognizing such connections, most landlords obliged their tenants to fallow fields every third year, and every other year if manure was unavailable. Reviving worn-out fields proved highly profitable-when enough land was enclosed. Under the banner of agricultural improvement, Parliament repeatedly authorized land enclosures that created large estates at the expense of common land, enriching the landed gentry and turning peasants into paupers.

English farmers gradually increased per-acre grain yields to well above medieval crop yields of twice the seeded amount, which were no greater than early Egyptian crop yields. Traditionally, historians attributed increased yields between the Middle Ages and the Industrial Revolution to the introduction of clover and other nitrogen-fixing plants into crop rotations in the eighteenth and early nineteenth centuries. Crop yields at the start of the eighteenth century were not all that much greater than medieval levels, implying that increased agricultural production came largely from expanding the area cultivated rather than improved agricultural methods. Wheat yields had risen by just a bushel and a half over medieval yields of ten to twelve bushels per acre. Yet by i8io yields had almost doubled. By i86o they had reached twenty-five to twenty-eight bushels an acre.