Ecological Agriculture Projects Logo

EAP Publications | Virtual Library | Magazine Rack | Search | What's new

Join the Ecological  Solutions Roundtable


Ecological Impact of Modern Agriculture

Judy Soule
Danielle Carré
Wes Jackson

We might call this the Age of the Recognition of Limits. A generation ago, the energy reserves that fueled the extractive economy were seen to be the primary limits affecting human progress in every endeavor, including agriculture. American scientists and policymakers alike were comfortable in their belief that these acknowledged finite reserves were large enough to buy the time necessary to bring the peaceful atom on line. It isn't that renewable resources were ignored. Americans did worry about the consequences of soil erosion and water pollution, of course, but there was a nearly full confidence in the technical fix for problems in the renewable sector of the economy. For agriculture, this meant that terraces, grass waterways, and proper incentives would take care of erosion. Simple solutions like indoor plumbing and chlorine in our drinking water would handle the water problem. And as for this nonsense about pesticides, Rachel Carson was just a hysterical extremist who was not an expert. The attitude about commercial fertilizers was that "nitrogen is nitrogen" and "plants don't care whether it comes from an anhydrous ammonia tank or from the nodule of a legume." And besides that, "We must feed the world." We had a sense of the heroic urge in us. It was also an era in which "hard-headed realism" was a favorite expression. But as author Wendell Berry has said, "A hard-headed realist is usually a person who uses a lot less information than is available." We have more information now than a generation ago, and the public is better informed.

The "problem of agriculture" is as old as agriculture itself, and although the core of the problem has always been soil erosion, new problems have been added. The high-energy epoch that fueled industry, which in turn made the industrialization of agriculture possible, simply exacerbated the old problems in agriculture and added more besides. And so we are dealing with an old subject with new data and the need for a fresh approach in our search for solutions. We will consider these problems one at a time even though they are inextricably intertwined.

Soil Erosion

Raindrops bombarding bare soil result in the oldest and still most serious problem of agriculture. The long history of soil erosion and its impact on civilization is one of devastation. Eroded fields record our failure as land stewards. W. C. Lowdermilk (1953) described several civilizations that collapsed because of erosion. At the end of his historical sweep of failed civilizations, Lowdermilk warned that unless we want to experience a similar fate, we had better heed these lessons of the past and safeguard our soils.

Soil erosion came to the nation's attention dramatically during the dust-bowl 1930s. Dust storms that blackened the sky were harder to ignore than the more severe problem of sheet and rill erosion due to rain. The Soil Conservation Service (SCS) was formed during this period under the energetic and imaginative direction of Hugh Hammond Bennett. The SCS encouraged planting shelterbelts of trees and grassed waterways and using contour ploughing and terracing. But by the mid-1970s soil erosion once again became a national concern. Farmers were encouraged both by the export potential and by Secretary of Agriculture Earl Butz to plant from fencerow to fencerow (Weasel, 1983), bringing 24 million additional hectares (60 million acres) into production (Crosson and Stout, 1983; Batie, 1983). By and large, the extra land was more prone to erosion. Larger machinery was purchased and pulled into the fields, and suddenly the terraces on sloping land and the windbreaks were nuisances (Weasel, 1983). Soil conservation meant increased costs to farmers, who were already burdened by the additional investments. The possibility of higher profits overshadowed any long-term benefits of conservation.

In 1977 the SCS, in an attempt to understand the amount of this precious resource being lost, began a new survey called the National Resources Inventory (NRI), published in 1981. The public was given a comprehensive review of sheet, rill, and wind erosion in the United States over a cropland base of 167 million hectares (413 million acres) (Crosson and Stout, 1983; Batie, 1983). An estimated total of 5.8 billion metric tons (6.4 billion tons) of topsoil was washed or blown away in the United States (USDA, 1980; Crosson and Stout, 1983). Some 2 billion metric tons of soil was lost from American cropland alone, an average of 15.4 metric tons/ha (6.8 tons/acre) (Crosson and Stout, 1983). Sheet and rill erosion were responsible for 1.7 billion metric tons being lost, while wind swept away 816 million metric tons (Crosson and Stout, 1983). The average "official" tolerable losses of 11.3 metric tons/ha (5 tons/acre) for deep soils and of 2.3 metric tons/ha (1 ton/acre) for shallow soils were violated (Crosson and Stout 1983; Batie, 1983).

Since erosion is not evenly distributed, in some areas the erosion rate greatly exceeded 15.4 metric tons/ha (6.8 tons/acre) In the Palouse Hills region of southeastern Washington and southwestern Idaho, where dryland farming is devoted to wheat, barley, beans, and lentils, there are numerous slopes ranging in steepness from 15 to 25 percent; here the soil erosion due to snow and rain was 113 to 227 metric tons/ha (50 to 100 tons/acre) (USDA, 1980). In southeastern Idaho, where hard red wheat is planted on slopes as steep as 35 percent, the annual erosion rate was 36 metric tons/ha (16 tons/acre) (Batie, 1983). Steep but fertile ground in southern Mississippi experienced a loss of 45 metric tons/ha (20 tons/acre) or more (USDA, 1980). Corn-belt losses up to 45 metric tons/ha were common (Batie, 1983).

Although the NRI was the most extensive survey yet produced, it did not give a complete picture of the amount of soil loss. The NRI relied on the universal soil loss equation (USLE) and the wind erosion equation (WEE) for estimating total soil loss (Batie, 1983). The USLE can estimate sheet and rill erosion, but it will not determine the amount of soil lost in gullies (Crosson and Stout, 1983). Wind erosion was determined for only 10 states in the 1977 survey (Crosson and Stout )

The more comprehensive NRI in 1982 documented an increase in the nat~on's cropland base from 167 million to 170 million hectares (413 million to 421 million acres) during the 5 years between surveys (Lee, 1984). The acreage in the land-capability classes and subclasses most subject to erosion increased disproportionately. The size of the highly erodable land-capability classes of V, VI, VII, and VIII increased 9 percent, while cropland in the high-quality classes I, II, and III increased only 1.2 percent (Lee, 1984). The sum of wind and water erosion averaged more than 18 metric tons/ha (8 tons/acre), or 3.1 billion metric tons (3.4 billion tons) on cropland. Forty-four percent of the cropland lost soil above the tolerance level (Lee, 1984). Nearly a fifth of the acreage was devoted to corn, a crop linked to high erosion rates (Bills and Heimlich, 1984).

Global losses are also high. Lester Brown (1981) reports that around 21 billion metric tons of soil is being lost worldwide. Even India and China, presently self-sufficient in food production, are losing soil beyond replacement levels. In India an estimated 6 billion metric tons of soil is lost annually from their croplands, representing a 4.7 billion metric tons/yr loss beyond the tolerance level (Brown, 1984).

Erosion is not always obvious, especially sheet erosion. Three steps characterize the erosion process: Soil particles are detached, then displaced by wind and rain, and finally deposited. Raindrops striking the bare soil can throw soil particles several centimeters. Sheet erosion, or overland flow, occurs when the water removes a relatively uniform thickness of soil. Rill erosion is easier to detect, since we can readily see the small channels, or rills, where water concentrates. The erosive power of the rill variety is responsible for a majority of our soil losses. (All from Morgan, 1979.)

Wind erosion also separates soil particles. Fine particles less than 0.2 mm (0.01 in) in diameter are transported high in the air over long distances (Lyres and Tatarko, 1986). Larger particles roll along the ground, become abraded, and add to the fine particles that are removed by the wind (Lyres and Tatarko, 1986; Morgan, 1979). Soil structure is damaged and soil texture altered as the coarser particles are left behind and organic matter decreases (Lyres and Tartako, 1986).

With such a wholesale altering of its basic structure, texture, and organic matter, the deposited soil no longer has the potential to produce its former yields (Rosenberry et al., 1980; Meilke and Schepers, 1986; National Soil Erosion-Soil Productivity Research Planning Committee, 1981). As the soil erosion continues, the subsoil becomes part of the tillage layer, and the fertilizer inputs and power requirements increase to the point where it is no longer feasible to cultivate the land (Rosenberry et al., 1983).

The problems of soil erosion are not limited to the farm. Soil from eroded cropland is the largest contributor to nonpoint pollution in the United States (Meyers et al., 1985). Sediment makes up the greatest volume by weight of materials transported in streams and rivers, and other pollutants (such as pesticides and fertilizers) can be carried along with it, either adsorbed on the soil particles or in solution (Clark, 1985). These pollutants adversely affect aquatic ecosystems either by directly affecting fish and other aquatic life or by altering spawning areas and food sources (Clark, 1985; Meyers et al., 1985). In addition, sediment accumulation is reducing the capacity of waterstorage facilities; approximately 1727 m3 (1.4 million acre-feet) of reservoir and lake capacity is filled each year with sediment (Clark, 1985). Dredging is a costly way of maintaining this capacity, and alternative reservoir sites are becoming difficult to find (Poster, 1985). The consequence is that municipal water sources are affected. The costs to clean the water are substantially higher since cleaning devices (sedimentation basins, filters, and chemical coagulants) need to be installed (Clark, 1985).

The USDA has promoted no-till and minimum-tillage practices as erosion control measures. In 1982, 4.7 million hectares (11.6 million acres) of farmland were in no-till, and the USDA has predicted that 85 percent of all cultivated cropland will be in some form of minimum tillage by the year 2000 (Hinkle, 1983). Unfortunately, no-till practices rely heavily on herbicides.

Some promising attempts to reduce the amount of soil erosion are the two conservation provisions of the 1985 Farm Bill (Wenzel, 1986). Under the first, the SCS will identify highly erodable land. Farmers who convert any highly erodable land to cropland will be denied such benefits as price supports and crop insurance (Wenzel, 1986), possibly keeping 13 million hectares (32 million acres) from being converted to cropland. The second provision, the Conservation Reserve, pays farmers to take highly erodable land out of production; it caused 3.6 million hectares (9 million acres) to be taken out of production in 1986 (American Farmland Trust, 1986). By 1986, Congress had a target of 18.2 million hectares (45 million acres) (American Farmland Trust, 1986), but the number is likely to go higher.

Future efforts at soil conservation should concentrate on developing farming practices that preserve and enhance the soil. Farmers should be subsidized to initiate soil-conserving practices, such as terracing and contour plowing. Also, we should look to natural ecosystems to provide our best example, such that soil surfaces are always protected by a canopy. This is difficult to obtain in an annual cropping system where the soil is tilled every year. Reduced-tillage practices have helped conserve soil, but the use of herbicides with this practice makes it undesirable in a sustainable agriculture. Perhaps in our search for new and more sustainable agricultural methods we will find other ways to conserve soil. Planting perennial cropping systems on sloping ground may allow continued use of this fragile land, or we may have to convert it to grassland and limit our cropping systems to flat bottomlands.


Adequate rainfall is never guaranteed for the dryland farmer in arid and semiarid regions, and thus irrigation is essential for reliable pro auction. Irrigation ensures sufficient water when needed and also allows farmers to expand their acreage of suitable cropland. In fact, we rely heavily on crops from irrigated lands, with fully one-third of the world's harvest coming from that 17 percent of cropland that is under irrigation (Poster, 1985). Unfortunately, current irrigation practices severely damage the cropland and the aquatic systems from which the water is withdrawn and partially returns.

Currently, the 255 million irrigated hectares (631 million acres) in the world use 70 percent of the total world water consumption (Poster, 1985). In the United States more than 20 million irrigated hectares (49 million acres), concentrated in the 18 western states and in the southeast, use 83 percent of the total water consumed (Fredrick and Hanson, 1982; USDA, 1985). Very few conservation efforts are currently devoted to water, probably because it has been regarded as a limitless resource. Farmers worldwide face the problem of limited supplies and the resultant competition over what is left. But even in the face of decreased supplies, the number of irrigated acres continues to expand. In the 1970s an additional 5.1 million hectares (12.5 million acres) per year were irrigated (Poster, 1985).

Water from underground aquifers supplies 40 percent of our irrigation needs in the United States (Pimentel et al., 1982). Groundwater provides an excellent high-quality source. In the United States the storage capacity of aquifers greatly exceeds that of rivers, lakes, and reservoirs (U.S. Geological Survey, 1984). But supplies are not limitless; aquifers recharge slowly from water percolating through the soil, and water in some deep aquifers can be considered a nonrenewable resource. The amount of water that is pumped from the aquifers exceeds the amount replenished in many areas of the world. In the United States approximately one-fourth of the groundwater extracted exceeds the replenishment (Stokes, 1983). As water tables fall, springs dry up and streams experience reduced flow (Poster, 1985; Pimentel et al., 1982).

The Ogallala aquifer is a good example of rapid aquifer depletion. This aquifer lies under parts of seven states: Texas, New Mexico, Colorado, Kansas, Nebraska, Wyoming, and South Dakota. It supports one-fifth of the U.S. irrigated cropland (Poster, 1985). The export push in the 1970s encouraged farmers to install extensive irrigation systems (Weasel, 1983). Farmers irrigated 3.2 million hectares (8 million acres) in 1978, compared to 0.85 million hectares (2.1 million acres) in the 1940s (Poster, 1985). A center-point irrigation system (the type now prevalent in this area) can use 3030 L (800 gal) of water per minute (Adams, 1981). Five hundred cubic kilometers (405.4 million acre-feet) of water has been withdrawn in the last four decades, and some hydrologists predict that given the current rate of drawndown, the aquifer will fail to yield in another 40 years (Ferguson, 1983). The exhaustion of the Ogallala aquifer is showing its effects; land in Texas, New Mexico, and Kansas is being taken out of production because of diminishing well yields and rising pumping costs (Ferguson, 1983). A total of 6 million hectares (15 million acres) in 11 states may be taken out of production because of water shortages; Texas alone may lose 1.2 million hectares (3 million acres) by the year 2000 (Stokes, 1983).

Declining water levels not only affect farmers. In the southeast, several cities depend on groundwater to meet their water needs. Tucson, Arizona, is a city completely dependent on an aquifer to meet its needs, yet only 35 percent of the water withdrawn is replaced by recharge (Stokes, 1983). Exhausted water resources will effectively curtail the expansion of these cities.

Land subsidence, the dropping of the land surface, is another major problem when aquifers are overdrawn (Bouwer, 1981). Numerous incidences of land subsidence have been recorded in California (Carbognin, 1985). Heavy pumping of the groundwater began after World War II and has continued to the present (Carbognin, 1985). In the 51 years between 1925 and 1977, 20 billion cubic meters (26.2 billion cubic yards) of land has dropped due to the extraction of 70 million cubic meters 57,000 acre-feet of groundwater (Carbognin, 1985). Groundwater withdrawal in Mexico City has caused 9 m (30 ft) of land subsidence, creating problems with their drainage system and with construction (Carbognin, 1985). In addition to subsidence, the depletion of aquifers along the coastlines of California, Georgia, and Florida has caused saltwater intrusions (Pimentel et al., 1982).

Depletion of streams and rivers is also a serious problem. River ecosystems require minimum levels of instream flow. Aquatic and riparian habitats are severely affected or destroyed when rivers are overdrawn (U.S. Water Resources Council, 1978). In the l8 western states most dependent on irrigation, 70 percent of the instream flow is often depleted (Fredrick and Hanson, 1982). For example, in the lower Colorado River more than 80 percent of the instream flow is depleted (Pimentel et al., 1982). River runoff may not be adequate for optimal fish and wildlife habitats in rivers of 14 western and southwestern states by the end of this century (Stokes, 1983). To protect these natural habitats, irrigation would have to be reduced to one-fifth of what it is now (Stokes, 1983).

Beyond depleting water resources, current irrigation systems can destroy cropland and reduce water quality. Waterlogging, salinization, and alkalinization have damaged about half of the world's irrigated lands (CEQ, 1980). Although this is only a small percentage of the world's total cropland, the losses are significant when the higher yields obtained with irrigation are considered. The soil and groundwater in arid zones typically have a high salt content, and natural drainage is usually poor (Kovda, 1980). Irrigation water drains slowly from fields, and as the water table rises, the soils become waterlogged in the crop root zone (Kovda, 1980). Continued application of irrigation water and seepage from unlined canals cause the salty groundwater to reach the surface (Kovda, 1980). As the water evaporates from the soil, a salty white crust forms on the surface, leaving the land unfit for cultivation unless expensive reclamation procedures are used. These problems are present wherever large acreages of cropland are irrigated (Kovda, 1977). Pakistan has enormous problems with salinization; 9.1 million of its 14.6 million irrigated hectares (22.4 million of 36 million acres) have some form of salinity problem (CEQ, 1980). In the San Joaquin valley about 162 million hectares (400,000 acres) are affected by high brackish water tables. Unless adequate drainage is installed, approximately 13 percent of this acreage will become unproductive (Poster, 1985).

Cropland salinity also creates water-quality problems in arid and semiarid river basins. Less than 50 percent of the water applied for irrigation is used by plants; the rest eventually returns to rivers and streams (Poster, 1985). The quality of the return flow is seriously degraded by salts, sediments, fertilizers, and pesticides. In addition, groundwater sources can become contaminated by the deep percolation of degraded irrigation water (El-Ashry et al., 1985). As a result, municipal and industrial water sources become polluted and users must pay higher water-treatment costs.

Applying large amounts of water in arid regions can also alter the ecosystem; in addition, such changes in farming practices as plowing, fertilizing, and monocropping (which usually accompany irrigation) also affect the ecosystem. Plant diseases, insects, and weeds, previously unable to withstand the dry conditions, can become serious pests (Ghabbour, 1977). The soil fauna, usually sparse in desert climates and adapted to drought conditions, can undergo drastic changes with irrigation; certain groups of soil fauna become dominant, destroying the previous balance. For example, earthworms, collemba, mites, and nematodes will invade newly irrigated soils at the expense of the original fauna unable to adapt to the new moisture levels (Ghabbour, 1977).

Human diseases have also increased with the widespread use of irrigation, especially water-transmissible diseases such as malaria and schistosomiasis (Obeng, 1977; Farid, 1977). The aquatic snail that transmits schistosomiasis inhabits sluggish waters common in irrigation projects, and schistosomiasis is now estimated to affect 200 million people in Asia, Africa, the Caribbean, and South America (Obeng, 1977). The disease has reached its current levels with the construction of irrigation projects. In the Gezira project in the Sudan, for example, the incidence of the disease increased almost 50 percent in a 20-year period after the expansion of the irrigation project in 1950 (Obeng, 1977).

Although irrigation in arid areas cannot be totally abandoned, we need to realize that water is a limited resource and that the arid ecosystems are fragile. Upgrading irrigation systems will ameliorate some of the problems outlined above. Instead of financing massive water-diversion projects, improving irrigation efficiencies can free water for other demands. Lining canals, reusing water runoff, and increasing the efficiency of flood and furrow systems can save enormous quantities of water. Unfortunately, the incentive to save water is lacking in the United States since farmers pay only one-fifth of the cost of the water they use (Ferguson, 1983). Limited supplies and rational costs will force changes in the allocation and use of water.

In addition, the type and extent of agriculture practiced in arid regions should be reexamined. Agricultural systems transferred from humid areas are not suitable. Investigating the unique qualities of arid ecosystems will help develop more appropriate agricultural practices.

Agriculture and the Loss of Genetic Diversity

As modern agriculture converts an ever-increasing portion of the earth's land surface to monoculture, the genetic and ecological diversity of the planet erodes. Both the conversion of diverse natural ecosystems to new agricultural lands and the narrowing of the genetic diversity of crops contribute to this erosion.

An estimated 852 million hectares (2.1 billion acres) worldwide have been converted to regular cropping or grazing in the past 120 years through deforestation, draining wetlands (or converting to rice), irrigation, and conversion to grazing (Richards, 1984; Salick and Merrick, Chap. 19 of this book). Although in North America from 1920 to 1978 the reversion of cultivated land to wild land about equaled the conversion to new agricultural land, elsewhere conversion has far exceeded reversion. In Africa 90.5 million hectares (223.6 million acres) were converted during this same period, and worldwide net conversion (conversion minus reversion) equaled about 420 million hectares (1.04 billion acres) (Richards, 1984). Current world population trends and the unequal distribution of power and resources will keep the pressure for continued conversion high (Wilson, 1985). In the United States the increased demand for exported food in the 1970s caused the conversion rate to climb dramatically; an estimated 24 million hectares (59 million acres) were converted to agriculture between 1972 and 1980 (Richards, 1984). Inevitably, some ecologically sensitive areas, such as the vital breeding grounds for wetland birds and waterfowl known as the "prairie potholes" in North Dakota, were victims of this agricultural expansion.

In the tropics, rates of land conversion to agriculture remain high (Richards, 1984), and there the impact on global species diversity is the greatest. The ongoing destruction of tropical forests is estimated to cause thousands of species extinctions annually (Wilson, 1985). Scientists estimate that 3 million species of plants (two-thirds of all species on earth) reside in the tropics, and about half of these are confined to tropical forests (U.S. Department of State, 1981; White, 1983). At the current rates of deforestation, one-fourth to one-third of all modern species will be gone in the next 30 to 40 years (U.S. Department of State, 1981; Batten, 1983).

The pressure on tropical forests is twofold. On the one hand, the need for foreign exchange, coupled with the industrial nations' voracious consumption of forest products, produces lucrative economic incentives for logging operations (World Commission on Environment and Development, 1987; Nations and Komer, 1983). Historically, economic factors have been the prime cause of most deforestation (Richards, 1984). On the other hand, inequitable land distribution has created a rapidly growing population of landless peasants. In Latin America a mere 7 percent of the landowners control 93 percent of the arable land. From Brazil and Colombia to Kenya, Thailand, Indonesia, and Madagascar the story is similar. Logging roads provide access to unclaimed lands, and once the valuable timber species have been logged out, the landless poor are encouraged to move in and clear the land. In the process, indigenous peoples are likely to be displaced, thus increasing the pool of landless poor. Typically, the cleared land will yield for only 5 to 7 years of intensive cropping, and then the peasants must look for a new plot. Often the newly cleared land is taken over by large landowners to raise export crops, such as beef in Latin America, at pitifully low productivity rates (World Commission on Environment and Development, 1987; Nations and Komer, 1983). Grazing cattle in cleared rain forest in Chiapas, Mexico, yields a mere 10 kg/ha (8.9 lb/acre) of meat per year. In sharp contrast, the traditional methods of the indigenous Maya yield 5896 kg/ha (5263 lb/acre) of shelled corn plus 4535 kg/ha (4049 lb/acre) of roots and vegetables per year for 5 to 7 years, then additional yields of citrus, rubber, cacao, avocado, and papaya for the following 5 to 10 years while the plots are allowed to regrow (Iltis, 1983). To remove the pressure on tropical forests, a complex set of changes must occur, ranging from (1) economic disincentives to the industrial nations' importation of tropical wood to (2) extensive agrarian reforms in the developing nations. These reforms include land redistribution that returns productive land to displaced peasants, the development and adoption of ecologically sound farming practices, and the goal of self-sufficiency in order to reduce the need for foreign exchange.

Modern agriculture's impact on the genetic diversity of crop plants is also profound. Recent gains in yield have been achieved by increased dependence on very few genetically narrow cultivars of only a handful of major food plants (Committee on Germplasm Resources, 1978; N. J. Brown, 1981; Mayer, 1981; Wilson, 1983; Duvick, 1984; Kannenberg, 1984). Only 10 to 20 crops provide 80 to 90 percent of the world's calories (N. J. Brown, 1981; Mayer, 1981)--or as W. L. Brown (1981) states, 15 species of cultivated plants "literally stand between man and starvation." As of 1970, in the United States 56 percent of the soybean crop, 71 percent of the hybrid. corn crop, and 41 percent of wheat acreage were occupied by only six cultivars each (Duvick, 1984). For red winter wheat, only two varieties made up 75 percent of the wheat land, with a single variety covering more than 50 percent of the acreage. Although the dominance of the top six cultivars was reduced somewhat by 1980--to 42 percent of soybeans, 43 percent of corn, and 38 percent of wheat (Duvick, 1984)--still, the genetic base of the varieties is limited (Kannenberg, 1984). The major cultivars of red winter wheat trace back to only two original cultivars: Turkey and Marquis. The hundreds of corn hybrids grown in the United States and Canada are largely based on about 12 inbred lines that originated from a few open-pollinated varieties of a single race (out of some 200 known races of corn). The U.S. soybean varieties were derived originally from six plants from Asia (N. J. Brown, 1981).

Throughout the world the attraction of high yields from new strains has caused farmers to abandon the more diverse, locally adapted varieties, or landraces (Committee on Germplasm Resources, 1978; Pluckiest et al., 1983; Van Sloten, 1984). For example, F1 hybrids have virtually replaced landraces of cultivated Brassica spp. in Japan, Korea, North and South America, and Northern Europe. China, the last holdout for genetic diversity of this group, is now losing landraces rapidly (Van Sloten, 1984). In the past 40 years, 95 percent of Greek wheat landraces have been lost because of the introduction of highyielding varieties (Pluckiest et al., 1983). Similarly, nearly all sorghum landraces disappeared from South Africa following the introduction of high-yielding hybrids from Texas (Pluckiest et al., 1983). Landraces of white potatoes, sugarcane, and tomatoes are already mostly lost (Committee on Germplasm Resources, 1978). The problem is compounded by governmental policy where official seed agencies emphasize uniformity and purity of seeds, as in the United States and Canada (Committee on Genetic Vulnerability of Major Crops, 1972). Solutions to this threat will need to be multifaceted.

Reduced genetic diversity in crops results in a loss of flexibility to meet future breeding challenges. Ironically, the landraces in jeopardy are essential to the maintenance of the high-yielding, genetically narrow cultivars that replace them (Committee on Germplasm Resources, 1978; Bertrand, 1981; Pluckiest et al., 1983; Smartt, 1984). Landraces are the breeders' sources of resistance to diseases and pests and of adaptation to climatic extremes and poor soils (Brady, 1981; Pluckiest et al., 1983; Kannenberg, 1984). For example, the source of powdery mildew resistance in California melons was a wild melon from India. Resistance to ripe rot in North American pepper plants came from a Peruvian species. High-yielding hybrid cucumber seed was obtained from a Korean strain (Bertrand, 1981). Wild genotypes of rice have helped to boost rice yields by an estimated $1.5 billion per year (Myers, 1983).

Several recent trends further increase the need for new sources of genetic diversity: Wide use of uniform varieties in monoculture has increased vulnerability to biological and physical stresses; expansion of agriculture into marginal lands requires special adaptation to climatic extremes, poor soil, and low water quality; consumer demands for appearance, flavor, and quality require continual innovation, as do such changes in farm technology as no-till farming, multiple cropping, and biological farming (Pimentel, 1977; Bertrand, 1981). Without the landraces and wild species to draw on in the future, breeders will reach the limits set by the variation available in the cultivars and have nowhere to turn for future innovation (Kannenberg, 1984).

Crop breeders have begun to respond to this dangerous loss of genetic diversity. Reduction in dependence on the top six cultivars of wheat, soybeans, and corn between 1970 and 1980 in the United States is one example (Duvick, 1984). Corn breeding has shifted most dramatically toward genetically broader varieties (Kannenberg, 1984). Pedigree breeding (where the best cultivars of each generation are the parents for the next generation), which results in a very narrow genetic base, is less popular now than it once was. By 1975 synthetic populations (intercrosses of several elite inbred lines), crosses of exotic and adapted lines, and population-based recurrent selection programs were producing 49 percent of the new lines of corn. Kannenberg (1984) advocates adoption of an even richer genetic foundation, which he achieves by a hierarchical open-ended system in which new genotypes are continually introduced into the breeding program.

Widespread recognition of the severity of the genetic erosion of crop species came in the 1960s, and by the early 1970s this recognition had turned to alarm among crop breeders (Williams, 1984). The Consultative Group on International Agricultural Research (CGIAR) was formed in 1971 to study the problem, and in 1973 the CGIAR created the International Board for Plant Genetic Resources (IBPGR). This board undertook the major tasks of collecting, organizing, and preserving seed sources representing as much of the worldwide diversity of crop and related species as possible. The funding for this program is international, and the seeds and information gathered are meant to be freely available to all. Although this program achieved a great deal in its first 10 years, especially in the collection of lines of the major grain crops, in 1984 the job was still far from complete (Committee on Germplasm Resources, 1978; Williams, 1984). Systematic ecogeographic searches in remote areas are still needed, with special attention to disease and pest resistance. Closely related wild species and most tropical crops have been largely neglected. Effective methods for preserving vegetatively propagated crops and those with short-lived seeds have yet to be found. Even though many seed sources are now safely preserved, most are still essentially unavailable to breeders because they have not been evaluated or characterized. This huge job will require major effort and funding. In many cases, in situ conservation of wild crop relatives and landraces is necessary to effectively preserve their genetic diversity, particularly for perennial vegetatively propagated crops (Ingram and Williams, 1984). The most direct way to accomplish the in situ preservation of genetic diversity may be to preserve indigenous subsistence agriculture. Where local varieties are found to be nutritionally superior, ecologically sound, and more reliable in unpredictable growing conditions, these systems should be maintained rather than replaced with modern, genetically limited varieties and farming techniques. .

Chemical Contamination

"In nearly all respects agriculture became an industry, sharing with the traditional manufacturing industries the problems of waste byproducts disposal." Young (1983) is referring in this quote to the changes in modern agriculture that have occurred over the past 45 years since the introduction of cheap inorganic nitrogen fertilizers. Six kilograms (13 lb) of pesticides, 63.5 kg (140 lb) of actual nitrogen, 19 kg (42 lb) of phosphate--these were the average chemical inputs per irrigable acre in California in 1980. In 1 year, California alone uses 55 million kilograms (121 million pounds) of restricted-use pesticides (Afford and Ferguson, 1982). Along with the promise of high production that these figures imply is the potential environmental pollution that these massive chemical inputs can cause.

Pollution of groundwater and surface water

Nitrate is currently recognized as the most serious agricultural chemical pollutant (Alfoldi, 1983). As nitrate shows up in drinking water, where it can cause methemoglobinemia in infants, attention is increasingly focused on nitrates in the groundwater source. Agriculture contributes to groundwater nitrate through the leaching of nitrogen fertilizer, animal wastes from feed lots, and organic matter from plowing under grasslands and crop residues (Guenzi, 1974; Aldwell et al., 1983; Alfoldi, 1983; Young 1983). Studies in Illinois from 1969 to 1972 showed that 73 percent of farm wells in areas with shallow aquifers had nitrate levels above the critical level considered safe to drink (Singh and Sekhon, 1978/1979). Reports of elevated nitrate levels in groundwater from England, Israel, Germany, France, and the Netherlands surfaced in the 1970s, and at a recent symposium other European countries, as well as Nigeria and India, joined the growing list (Young, 1983).

Where water tables are shallow, nitrate can reach the water table quickly (Zaporozec, 1983). A number of studies have shown correlations between nitrogen fertilizer use and polluted groundwater (Singh and Sekhon, 1978/1979; Zaporozec, 1983; McWilliams, 1984). In addition, irrigation on porous, shallow soils exacerbates the problem of nitrate leaching. An experiment conducted in Minnesota on sandy glacial outwash soils above a 4.5-m (14.8-ft) water table clearly demonstrated an increase in nitrate in the groundwater after onetime applications of nitrogen fertilizer (Gerwing et al., 1979). Deeper water tables can also become contaminated by nitrate, but fewer cases are known. In a 1966 survey of thousands of rural Illinois wells, only 1.4 percent of the water samples from deep wells [deeper than 30 m (98 It)] exceeded the Environmental Protection Agency (EPA) limit of 10 mg/L (38 mg/gal), as compared to 23 percent of the samples from shallow wells. But in 1982 in rural Rock County, Wisconsin, 10 percent of the water samples from deep wells [deeper than 46 m (151 ft)] exceeded the limit (Zaporozec, 1983). In England, high nitrate concentrations were found to a depth of 35 m (115 ft) below the minimum water table (Young, 1983). The movement of nitrate into deep groundwaters occurs so slowly in relation to horizontal groundwater movement that it is difficult to prove a direct association with fertilizer applications (Singh and Sekhon, 1978/1979); even so, nitrate levels are usually correlated with overlying land use. One study found the highest nitrate concentrations below cattle feedlots and irrigated alfalfa, followed by fallow wheat and then by virgin grasslands, which had the lowest concentrations (Singh and Sekhon, 1978/1979). Decontamination will take place as slowly as contamination. Once below the root zone, nitrate is highly stable (Singh and Sekhon, 1978/1979) and can percolate slowly toward deep aquifers for years. Even if all fertilizer application were suspended immediately in areas where deep aquifers are beginning to show nitrate pollution, the inputs of past years would still slowly percolate toward the aquifers, and nitrate input to the groundwater would continue for many years.

Deep aquifers are fairly well protected from pesticide contamination (Alfoldi, 1983), since all but the most persistent pesticides degrade before they reach a deep aquifer. Shallow aquifers, however, are more commonly contaminated with pesticides. In the shallow aquifers of Wisconsin's Central Sands region, the pesticide aldicarb has reached the groundwater in amounts above the EPA limit (McWilliams, 1984). In California, agricultural use of DBCP, a highly leachable nematicide used on a variety of crops, contaminated wells until its use was suspended in 1977 (FAO, 1979; Alford and Ferguson, 1982; Lecos, 1984).

Surface waters receive agricultural chemicals by runoff from field applications and feedlots and from the dumping of excess chemicals (Afford and Ferguson, 1982). Phosphorus from fertilizer applications binds to soil particles, whereby it can be exported in erosive runoff. Soluble organic forms in animal wastes or plant residues can also move into surface waters. Phosphorus is often the limiting nutrient in aquatic systems and is thus a major contributor to eutrophication (Oglesby, 1971). Nitrate, which is highly soluble and easily transported in runoff, also contributes to eutrophication (Oglesby, 1971). In combination, these two fertilizers have contributed much to the degradation of streams and other bodies of water in agricultural regions.

Pesticide runoff is normally low, with losses`of 0.5 percent or less (Wauchope, 1978), giving concentrations of 1 rLg/L (3.8 g/gal) or less (Guenzi, 1974). Under the worst conditions--a major rainfall or heavy irrigation within 3 days of application, the most mobile pesticides, sloping ground, and fine-textured soils--the losses may reach up to 15 percent; however, under these conditions the pesticide is likely to be in low concentration (Wauchope, 1978). Higher concentrations of pesticides in runoff occur when the runoff volume is low; even so, the concentrations are usually below the acute toxicity levels for aquatic organisms (Guenzi, 1974). Ironically, contaminated runoff is less likely to poison aquatic life than to damage other crops. Irrigation water contaminated with herbicides in one field can reduce crop productivity in another field. Picloram concentrations as low as 0.1 ppm can reduce sugar beet yield significantly (Guenzi, 1974).

The dangers of persistent pesticides in aquatic ecosystems, especially the organochlorines (e.g., DDT), have been known for many years. These chemicals, although not highly toxic at low concentrations, stay in the bodies of consumers and thus accumulate up the food chain (A. W. A. Brown, 1978; Guenzi, 1974). Top predators (such as fish-eating birds) consume toxic dosages with disastrous results. Agricultural use of toxaphene from 1957 to 1961 in the Klamath basin of northern California killed white pelicans and other fish-eating birds in wetland refuges (A. W. A. Brown, 1978). As these consequences became known, and as many organochlorine compounds came to be banned in the richer countries, they began to be replaced with shortlived, highly toxic organophosphates (such as parathion). Although toxic in minute quantities, the short life of such chemicals makes them unlikely to accumulate in bodies of water, and it also makes them difficult to study (Alfoldi, 1983; Guenzi, 1974; Wauchope, 1978). So far, the presence of nonpersistent pesticides has been documented, but no evidence of permanent impact in aquatic systems has been found (Wauchope, 1978). Studies set up to determine the impact of these pesticides have been limited to the direct application of pesticides to streams at higher concentrations than those found in agricultural runoff. These studies show that aquatic organisms vary widely in their sensitivity to different pesticides. Studies of organophosphates have shown that stone flies are sensitive to malathion, but less so to trichlorfon. In one study, malathion caused only a temporary depression of stream insect fauna, whereas fenitrothion caused extensive kills of stream insects (A. W. A. Brown, 1978). Species-specific responses to particular chemicals have been used to infer possible damage from natural agricultural runoff. A stream community in North Carolina that was affected by agricultural runoff showed a pattern of species reduction that matched the pattern caused by the direct application of carbaryl to streams in other studies--a correspondence that led Lenat (1984) to infer that this pesticide may have been present in the runoff. So far, studies of the effects of long-term, low-dose exposure of aquatic systems to nonpersistent pesticides are rare to nonexistent. Such studies should be pursued in order to ensure that we avoid serious environmental consequences.

Residues on food

Pesticides also pose hazards by direct contamination of foodstuffs. To what extent is our food contaminated with pesticide residues, and how much of a hazard is this?

The long-lived pesticides, such as DDT, other organochlorines, and parathion, are most likely to leave persistent residues (Bull, 1982). Even though most persistent pesticides are banned in the United States and other wealthy countries, they are still persistent where previously used, are still used elsewhere, and show up on foods. Others, such as organophosphates, tend to break down so rapidly that they are unlikely to show up on food unless applied to crops very close to harvest time, but they are often more acutely toxic. Approved pesticides leave little residue on crops when used according to directions. National and international standards of acceptable residue levels are based on approved usage and maximum acceptable daily intake (MADI) levels; thus, theoretically, foods should be safe. However, it is not practical to monitor all crops and shipments, residues are not always detected even when tests are done, and pesticides are not always used according to directions; thus, dangerous residues do make their way to the foods that people eat.

In the United States, public concern centers on the residues on imported foods (Lecos, 1984). Although the Food and Drug Administration (FDA) maintains an active monitoring program that makes it advantageous for growers to comply with U.S. standards, only a small percentage of shipments can actually be tested. According to FDA figures, some 2000 food shipments crossing the Mexican border are tested annually (Thompson, 1984). This is indeed a small percentage of the Mexican import traffic, which exceeds 900 truckloads per day during the peak of the January-to-April shipping season. In 1986 the FDA checked nearly 8000 samples, less than 1 percent of the imported shipments of fruits and vegetables. Rather than hold perishable produce at border checkpoints for the average 18 days it takes for the test results, shipments are routinely allowed to pass (GAO, 1986b). Positive tests, indicating dangerous pesticide residues, result only in the stoppage of future shipments or in recalls, if the food can be traced. From 1979 to 1985 an average of 6 percent of the shipments sampled were found to have illegal residues--a rate about twice the average rate of contamination of domestic foods. Online little more than half (55 percent) of the illegal shipments were prevented from reaching consumers, and in only a few cases were damages collected from offending shippers (GAO, 1986a; Bull, 1982).

How often pesticide residues actually turn up on the table in the United States is largely unknown. The FDA's total-diet studies, conducted once a year in four regions of the country, reveal that pesticides do show up frequently but are "consistently" below the limits set by the Food and Agriculture Organization (FAG) and the World Health Organization (WHO) (Farley, 1988). DDE, a breakdown product of DOT, was still found in about 40 percent of the food sampled in 1983. This chemical, along with other organochlorines, has declined steadily in food since the 1970s. At the same time, organophosphates have increased, corresponding to their increased usage. Malathion was found in about 20 percent of the food sampled in 1983. Increases in pesticide residues found in potatoes and peanuts in the early 1980s stemmed from the greater use of sprout suppressants and soil fungicides, respectively (Penner, 1984).

The EDB scare of the early 1980s is an example of a case where residues went undetected for a long time. EDB (ethylene dibromide) had been used since 1948 as a soil fumigant and as a postharvest fumigant on citrus and grains. Although a suspected carcinogen, EDB was thought to leave no residue, but the sensitive tests developed in the 1970s found that it did leave a residue. In 1983 it was found in cake mixes, cereals, and other foods from grains, and it was found in groundwater in several states where it had been used as a soil fumigant.

Public reaction was so strong that in September 1983 the EPA issued an emergency ban on the use of EDB as a soil fumigant. Bans on its use on grains and citrus followed in early 1984 ("Pesticide Is Banned," 1983; "New Moves," 1984; "EPA to Limit," 1984; Thompson, 1984). Although our food supply may now be nearly free of this dangerous chemical, over 30 years of use left a legacy of contaminated water and untold increases in cancer risk for the generation of consumers and farmers raised with it ("EDB's Long-Lasting Legacy," 1985; Barles and Kotas, 1987).

In third-world countries, pesticides are commonly misused because information on their proper use is not generally available. Their labeling is poor and often not in the native language, extension advisory services are usually unavailable, and proper usage outside of the temperate zone may be unknown, even by the manufacturers (Bull, 1982). In one market in India, a 7-year study showed that organochlorine residues were above the tolerance limits in more than 35 percent of the food. Forty percent of the samples of horticultural crops in Kenya contained excessive residues. A 1978 study. turned up rice samples from Sri Lanka that contained excessive residues of dieldrin and heptachlor. While exported foods are usually monitored carefully so that they will pass the inspections that the richer importing nations impose, locally marketed foods are usually not monitored, yet these are most likely to have residues of the short-lived, highly toxic organophosphates because of the short time between harvesting and marketing.

The responsibility for correcting this problem seems to lie with the richer nations of the world, which have profited by the export of pesticides and the technology of modern agriculture at the expense of the citizenry of third-world nations.

Pesticide resistance

In addition to directly poisoning our environment and our food, pesticides pose a serious threat to our food-production system itself. From one viewpoint, pesticides are the wonder chemicals that have increased food production by 20 percent since 1940 by reducing pest damage. Yet over the same period they have also created at least 261 strains of insect species, 67 strains of plant pathogens, 2 strains of nematodes, and 4 (by some counts 19) strains of weeds that they cannot kill (Conway, 1982). While insecticide use increased tenfold since the 1940s, crop losses to insects doubled (Pimentel et al., 1983). The key to this paradox is the strong selection for resistance that pesticides exert on their target pests. Pesticides never kill 100 percent of a pest population, and the survivors tend to have a lower susceptibility to that particular chemical. With every repeated application of the same pesticide, these naturally resistant individuals make up a higher percentage of the population, until a highly resistant strain of pest evolves. When the conditions are right--the pesticide kills a large percentage of the pest population, the pest completes several life cycles per year, and little immigration from untreated populations occurs--resistance can develop very rapidly.

Resistance to one pesticide often confers resistance, or faster development of resistance, to a whole family of related pesticides (Bull, 1982). Alternating different pesticides or applying a mixture of chemicals can sometimes delay the development of resistance, but it can also promote the development of superresistant pests. Superpests, resistant to multiple pesticides, have already developed and threaten a number of crops throughout the world. So far, widespread famine has not resulted from superpests on food crops, but regional crises have occurred and some larger problems loom ahead. In North and Central America, cotton production has the worst record of superresistant pests. The cotton bollworm and budworm are both resistant to organochlorines, organophosphates, and carbamates and are becoming resistant to the new pyrethroids (FAO, 1979; Conway, 1982). Worldwide, 33 species of cotton pests were resistant to pesticides as of 1982, an increase from 20 in 1965. Rice, the staple food of a large proportion of the world population, now has 14 pest species resistant to one or more insecticides. Without pest control, experts estimate that 50 percent of the rice crop in Japan, India, and the Philippines would be lost to pests (Conway, 1982). The diamondback moth, which feeds on brassicaceous crops, has developed multiple resistance to at least 11 insecticides, including some from all the major groups, including pyrethroids (Bull, 1982). This pest now threatens the production of these crops in southeast Asia (FAG, 1979; Conway, 1982). Farmers there may spray three times a week and use 11 different insecticides in the course of a season, often spraying less than a week before harvest and in higher-than-recommended doses (Bull, 1982).

Pesticide resistance is considered inevitable, with no type of chemical immune (A. W. A. Brown, 1978); formerly effective pesticides cannot be successfully reintroduced (Bull, 1982). In order to avoid major pest-caused crop failures in the near future, we cannot simply rely on chemists developing new chemicals. Rotation of pesticides may delay the development of resistance in some cases, but these cases will require careful, regionally specific tailoring, and the research has yet to be done (Conway, 1982). Under normal field conditions, tolerating low kills slows the rate of resistance development because of the lower selection pressure exerted. Any increasing migration of pests from untreated areas into treated areas also slows the evolution of resistance (Conway, 1982). Both lower kills and higher migration probably mean that farmers would have to tolerate greater crop damage by pests. Integrated pest management is one approach to a longer-term solution, but it still may be vulnerable to the development of resistance. The development of pest-resistant crop strains, a major thrust of crop breeding programs, suffers much the same malady as the pesticide solution; it becomes a race between the breeder and the pest as to which will develop resistance first.

Rather than looking at ways to kill pests without exerting selection pressure for resistance, we might do better to turn the question around and ask, How can we successfully do agriculture without having to kill pests? Perhaps the cropping systems themselves can be altered to minimize pest population buildups in the first place. Traditional crop rotations and mixed-cropping systems hold potential for pest control by suppression rather than by killing (Pimentel, 1977). Development of new perennial crops, drawing on the greater natural resistance of perennials, holds great promise (Jackson, 1987).


Despite the above chronicle of ecological disasters that attend it, agriculture is still potentially a renewable enterprise. In a certain sense, nothing is new here. In every century for 10,000 years on almost every agricultural acre in the world, ecological capital has slipped seaward from its wilderness-built home or been degraded by toxic materials. Still, on a global scale, agriculture is seen as potentially renewable and therefore as fundamentally different from the industrial sector of society. It is only in the last 50 years, with the expansion of industry and the chemicalization of agriculture, that the inherently extractive economy has acted as though the renewable resources that support agriculture are fair targets for exploitation in industrial terms. That is what makes the modern era different. That is what makes the current agricultural economy more brittle than almost any agricultural economy in history. It is hard for us to see this, perhaps, because it is hard for us to imagine our energy and aquifer mines lying hollow. It is hard for us to believe that our well water and air, our pleasingly packaged food, and our perfect produce could contain invisible poisons. It is hard for us to grasp the value of genetic diversity or to foresee the consequences of rampant extinctions. It is hard for us to believe that we may one day come to the end of our magical ability to produce everhigher-yielding crops. It is hard for us to comprehend the total loss of our vast tropical forests or to anticipate the climatic changes when the earth's belly is belted in barren, sterile soils rather than the green, moist vegetation. It is hard for us to imagine a world where ancient salts sterilize the land and young chemical pesticides and fertilizers lie below our agricultural surfaces like demons in quiet prisons of degraded soil.

The outlook is not entirely bleak, for solutions to all these problems lie in lessons we have learned and can still learn from nature. If we turn our attention away from the extractive industrial model and begin to focus on nature's models of productive ecosystems as our guide for agricultural systems, we may yet see truly sustainable agriculture emerging. It isn't that nature learns faster than humans. It is just that she has been at it longer.


Adams, J. L. 1981. "Pies on the Prairies," Frontier, May, pp. 15-18.

Aldwell, C. R., D. J. Burdon, and M. Sherwood. 1983. "Impact of Agriculture on Groundwater in Ireland," Environmental Geology, 5:39-48.

Alfoldi, L. 1983. "Topic 2: Movement and Interaction of Nitrates and Pesticides in the Vegetation Cover-Soil Groundwater-Rock System," Environmental Geology, 5:19-25. Alford, H. G., and M. P. Ferguson (eds.). 1982. Pesticides in Soil and Groundwater: Proceedings of a Conference, Division of Agricultural Sciences, University of California, Berkeley.

American Farmland Trust. 1986. "Confusion Hampers Conservation Reserve," American Farmland, 6(5):1.

Barles, B., and J. Kotas. 1987. "Pesticides and the Nation's Ground Water," EPA Journal, 13(4):42-43.

Batie, S. S. 1983. Soil Erosion: Crisis in America's Croplands? The Conservation Service Washington, DC.

Batten M. 1983. "The Rush Is On to Study Jungles," International Wildlife, 13(3):16-19.

Bertrand, A. 1981. "Biological Diversity: Basic for Agricultural Success,"

Copyright © 1990 McGraw-Hill. Excerpted from Ecological Impact of Modern Agriculture, pp.165-188.