The Natural History of the Present, Chapter 15

Chapter 15: Fields

Farmers and loggers are the landscapers of the continent. The 900 million acres of the United States (about half the land area) that is cropland, rangeland and pasture is managed by less than two percent of the population. Farmers and loggers shape the larger landscape and determine its ecological effects: its vegetation, its rates of erosion, its downstream, downwind, down-through-time effects. Industrial chemistry also shapes the landscape, though more subtly, at least to the casual observer: these changes show up in changing patterns of abundance of plants and animals. For the most part, as some species of birds and trees disappear, others take their place; the forest (perhaps less stable, or less useful) still has birdsong and still looks green. Developers also do their part, especially near rivers and coasts, where most people live. Changing patterns of terrestrial land-use propagate downstream through riverine marshes and wetlands and into the oceans, influencing the productivity of bays and estuaries, and thus of marine and estuarine fisheries. Seventy percent of marine fishes spawn in estuaries; and eighty to ninety percent of fish catches occur on continental shelves. Because of this, intensive crop production in upstream river valleys, with the resulting erosion, can reduce the total amount of food a landscape produces, through agriculture’s effect on the associated riverine and marine fisheries.

Much of the American landscape is settled in an anachronistic pattern. Buildings and industrial installations are set close to surface waters; or to ocean harbors. Roads follow river terraces, the paths of least resistance; these landscapes were carved out by glacial meltwaters and leveled by subsequent downcutting by the river and accumulation of streamside sediment: prepared by river geology and time. Rivers were also useful for floating logs, for running water-powered mills, as sources of water for people and animals. Some rivers floated boats, which were smaller in earlier centuries than now; large birch canoes (the “North Canoes” of the Hudson’s Bay Company) held several thousand pounds of furs, worth paddling halfway across a continent. Heavy loads could be floated more easily than drawn by wagon or carried on foot. Later, canals followed rivers, but by the middle of the nineteenth century, railroads had made river and canal transportation obsolete. Only public funding keeps it alive.

Labor, whether human or mechanical, lets one ignore topography. The cut of a nineteenth century railroad drives west through the Pennsylvania countryside, rising slowly above sealevel, its right-of-way dug out by hand. It runs almost entirely below the surface of the ground: the point was speed. The towns along the little Adirondack river where I live cluster along the water, about six miles apart. Three miles is an easy trot by horse, a moderate walk for a man. The towns were once trading settlements for farmers in the floodplain and up the side valleys, now for the most part they are simply sites of human habitation, property whose value is found in its remoteness, as automobiles and modern roads (artifacts of machine labor) have made most of the previous economic infrastructure obsolete. People shop in supermarkets thirty or forty miles away (stores whose headquarters are in Los Angeles or Holland), and children are bused to schools located sometimes in a village, sometimes merely centrally, that is, economically speaking, nowhere. Trucks made log drives on the local river obsolete (good news for the fish); the rich farmland of Ohio, and the irrigated vegetable fields of California (their water provided at 2% of its real cost by the government, in the interests of settling the landscape), together with refrigeration and railroads (also government supported), made much of Northeastern agriculture obsolete.

We live in a new world: still with abundant energy from fossil fuels, and still with remnants of the original, pre-agricultural ecosystems that once covered the planet. Such ecosystems are relatively recent as wholes; especially in the temperate zone, they are post-glacial, post-modern-climate, or approximately 6 to 15 thousand years old, depending on the place. These ecosystems were assembled quickly because they were assembled out of much older pieces. Some of these pieces now don’t work very well; for instance the North Temperate trees that now occupy the Mediterranean basin shed their leaves during the mild and damp Mediterranean winter and then leaf out and grow during the Mediterranean summer drought. These trees colonized the area as the glaciers retreated; and as the climate became more and more dry, they became less and less well adapted. They still work, but would be out-competed by better adapted vegetation, transplanted from someplace else. Similarly the American Southwest lacks species of browsers that could make use of its difficult vegetation; and forests throughout the Americas lack large mammalian fruit-eaters and tree-breakers (the elephants and their relatives) and must rely on birds to eat their large fruits and fire to renew the forests. A diverse mix of browsers and grazers would produce more animal flesh per acre in the dry rangelands of the Southwest than cattle alone: should the elephant, the camel, the Chacoan peccary, the llama, panther, lion and jaguar (all once resident here) be reintroduced to North America? Instead we introduce into such alien landscapes the crop plants and animals of the ancient Middle East, including our house cats: our current civilisation.

The world turns out to have larger limits. (It is not after all so subdueable.) It has limits in terms of the gases its atmosphere can absorb, such as the excesses of the greenhouse gases carbon dioxide, nitrous oxide and methane. These gases are produced by the combustion of fossil fuels, the burning of forests and grasslands, agriculture, domestic animals, land development, oil drilling and coal mining, as well as by lightning, beaver wetlands, lakes and digesting termites. It has limits in terms of the compounds of chlorine and bromine that interfere with the stratospheric ozone layer. (Nitrous oxide from agricultural fertiliser also has a place here.) This layer shields the green world from ultra-violet light from the sun. Chlorine compounds as a whole have increased in the stratosphere by 10 times since the 1950s; one reason the ozone hole keeps growing despite the phasing out of the worst of the offending chemicals is that new, useful chlorines and bromines are constantly invented; and people resist giving up old ones, such as the agricultural fumigant methyl bromide, used to sterilize soils in which crops like strawberries are grown. (In most such cases other, sometimes less profitable substitutes exist, such as soil sterilization by sunlight or crop rotation in the case of strawberries.) There are limits to what the green world can absorb in atmospheric aerosols of sulfur and nitrogen, in heavy metals, in hormones or hormone mimics in sewage water, in mutagens, or chemicals that attach to DNA; limits to the amount of carbon dioxide the oceans can absorb without reducing the ability of their plants and animals to make calcium carbonate shells (necessary for organisms like corals, clams, some algae); limits in what new plants, animals, insects, fungi can be spread among the different continents and seas without seriously simplifying existing ecosystems; limits on how much an ecosystem can be broken up and still function. Roads link the industrial landscape. Through roads it penetrates the countryside. Every road has its effects: a minor one is that songbirds avoid road noise; and those that nest near roads have smaller clutches. (Perhaps, like people, in whom road noise causes sleeplessness, compromised immune systems, and a slightly elevated risk of heart attack, they find road noise stressful.) Chemicals that end up on roads include rubber dust that contains polycyclic aromatic hydrocarbons, zinc, and cadmium; brake-lining dust containing copper, nickel, chromium, and manganese; lubricants containing cadmium, copper, molybdenum, vanadium, zinc; motor oil; brake fluid; antifreeze and unburned hydrocarbons from fuel (the last two, like cadmium and the aromatic hydrocarbons, are cancer causers and mutagens). Phosphorus and nitrogen also run off roads, in surprising amounts. (Nitrogen levels in national forestland near Denver are 10 times those in more isolated forests. Much of the nitrogen is thought to come from automobile traffic, that is, from the burning of gasoline.) Much material in road runoff could be contained in properly maintained stormwater ponds, though proper maintenance would be labor-intensive and expensive. (Recycling the metals and nutrients might help pay the costs: for instance, the concentration of platinum in road dusts would make mining them profitable.) Roads interfere with natural drainage patterns, flooding areas above them; they interrupt subsurface flows and lower water tables (roads are an efficient drainage mechanism); they change subsurface flows into connected streams; they isolate animal populations: a population of bank voles in southern Germany separated by a fourlane highway cross it so seldom that the animals on each side are becoming genetically distinct. Grizzly bears, especially females, also hesitate to cross roads, with the same effect. But roads also provide habitat: butterflies sun on them; moths fly along them; bats hawk for insects above them in the evening; migrating woodcock and robins search for earthworms in thawed roadsides in spring; flickers forage for ants; ravens cruise for roadkill; winter finches pick up grit; young loggerhead shrikes seek beetles near their edges and are run over by cars (one reason the species is in such steep decline). Turkey hens bring their poults to forage for grasshoppers in roadside edges; singing amphibians are crushed on wet nights (some populations which migrated across roads to breed have been exterpated); alien plants and animals follow roads to new environments. The drying effect of roads extends a hundred yards into the forest, as does the effect of salt spray. Invertebrate variety is reduced in pools near northern roads, probably because of the use of road salt. A million vertebrate animals a day may die on North American roads (or 10 times that), and untold numbers of insects: a writer has compared cars to filter feeders cruising through the insect world. Thus roads extend the industrial world into the countryside.

Despite what another writer has said, we have not reached the end of nature. We will not reach the end of nature until we are all dead. What is happening is that the natural world we are attached to is threatened. Many of its larger animals are going to go extinct, and populations of its less charismatic members continue to collapse (such as the amphibians, whose bodies form the base of many terrestrial food webs). The natural world our economic system is constructing will be less abundant, less joyful and work less well. The costs of its decline, in terms of ecosystem degradation and climate change, may amount to 20% or more of world economic product. We can probably deal with this. Much will depend on how we handle the transition to a poorer earth. Our success as an animal (biological success is measured in numbers, and humans are the most abundant mammal in their size class ever to inhabit the planet) has brought our influence on the planet to levels of long-term geological processes, of climate, of cataclysmic events, of the processes dominated by micro-organisms. The earthly landscape dominated by micro-organisms (“slimeworld”) endured the longest. It lasted from about 4 billion years ago to 500 million years ago, when the metazoans (multi-cellular animals) evolved and ate up all the available slime. Slimeworld retreated to underground rocks, to deep ocean muds, anoxic wetlands, water surfaces, to soils and the atmosphere, and remained greatly reduced on the earth’s surface, which was taken over by green plants. But the archaea and bacteria of slimeworld still play the defining role in making the earth habitable: in maintaining oxygen levels; in manufacturing cloud-seeding chemicals; in fixing and unfixing nitrogen gas into forms plants and animals can use. Modern trawls of genetic material of the ocean’s surface indicate there are thousands of species of unknown bacteria present in very small numbers. Such organisms, now just hanging on, are ready to expand in numbers when conditions change. Interferring with processes on this level is more risky for us; at the least, real estate values fall as lakes overfertilized by septic tank effluent or agricultural runoff are taken over by algae and cyanobacteria, eutrophicate and stink.

The fishy worlds felt the effects of agricultural settlement first and strongest. More extreme environments would join them in first showing the effects of industrial chemistry: mountain ridges, where ultra-violet light is strong and where industrial aerosols are filtered out of mountain fogs by tree needles and accumulate in soils (mountaintops in Central America concentrate the pesticides used in the lowlands, which rise on the heated air and condense in fogs; levels of pesticides on mountaintops can be 10 times those in the fruit plantations below, perhaps one reason for the decline of amphibians in Central American cloud forests); the Arctic, where thanks to atmospheric distillation and the constant movement of warm air north, industrial chemicals tend to end up. Fish swim in the denser fluid of water, which also accumulates chemicals running off the ground, condensing out of the air, concentrating in the surface layers of lakes and oceans; fish accumulate toxins through their food and absorb them directly through their skin and gills. Starting three and a half centuries ago in the United States, mill dams stopped fish spawning runs; and stocks that had formerly spawned above the dams died out. Sawdust from sawmills built over rivers covered spawning beds. Sawdust spread out over the bottom of Lake Michigan for several miles from Green Bay, Wisconsin; removing sawdust was always a problem, and running water removed it. Silt eroded from fields; and later, industrial effluent, oil, stockyard waste, and sewage made rivers less and less habitable for fish. But dams, the muddying and warming of rivers from agricultural erosion and overfishing, reduced fish numbers long before pollution became a problem. Dams, overfishing and agriculture (the destruction of riverine habitat), spelled the end of the Atlantic salmon in both North America and Europe. The Atlantic salmon’s population is now at less than 1% of historic levels, and there are now 900 dams on New England rivers that once held salmon. Cores from Chesapeake Bay show changes in its diatom population to species that tolerate increased turbidity as early as 1700, a hundred years after settlement began. The increased turbidity was caused by Europeans growing corn and tobacco in the watershed. Both tobacco and corn are demanding crops and required the constant clearing of new forestland. Trees would be girdled and tobacco or corn planted for three years; after that wheat might be grown for two years; then the land would be abandoned. Recovery took about 30 years. Tobacco planters sent their product out by ship (most plantations had a wharf on a river or the Bay itself) and so were sensitive to the dangers of siltation caused by agricultural erosion. At first, tobacco (the cash crop, worth six times any other and shippable) was grown by slaves or indentured servants using hoes on only the best, least erosive land.

Chesapeake Bay illustrates the effects of intensive settlement on the ocean edge. Chesapeake Bay is the drowned channel of the Susquehanna River. It was created by the rise in sea level at the end of the last ice age. The shape of the bay was originally hollowed out by a comet. The pre-settlement bay was one of the most productive estuaries in North America. One of its keystone species was oysters, which by filtering the water as part of their feeding behavior kept the water column clear. In 1600 oysters grew on reefs built up of oyster shells that had accumulated over the last 4,500 years. Sea levels had risen 3 feet per 100 years from 11,000 to 8,000 years ago, then slowed to 6 inches per 100 years by 4000 years ago. The more stable estuarine conditions let the oysters and other bay flora and fauna establish themselves and provided the resident peoples with abundant, dependable acquatic resources. Oyster reefs formed a rough, irregular shelf platform along the edge of the bay and its tributaries, protecting the shoreline and extending out to a depth of 6 meters (about 19 feet), which corresponds with the layer of rapid increase in salinity and therefore of density of the water column in spring and summer. The reefs, formed up to mean low water, broke the surface at low tide. Chesapeake natives fitted their canoes with a shallow keel to protect them from the sharp shells. Water draining off the reefs may have increased the vertical mixing of the water column in spring and summer, oxygenating the deeper waters more completely and reducing the vertical variation in salinity, letting the oysters extend their domain to greater depths. Native Americans began intensive use of the biotic resources of the bay about 5000 years ago. Oyster harvesting increased 3000 years ago. Signs of large scale drying of fish and shellfish appear 1000 years ago.

Oyster reefs make life easier for young oysters. After oysters spawn, and the fertilized eggs hatch, their larvae float about in the sea until it is time for them to attach and grow into adults. Attachment is the one big thing a larval oyster has to do. An oyster shell offers an oyster larvae a firm place to attach; and the young oyster uses the dissolving calcium in its ancestor’s shell to build its own. Oyster reefs are home to hundreds of other species, including mussels and sponges. Many of these species are fish, food for fish, or food for organisms that are food for fish. In the 1600s, European ships navigated around oyster reefs, whales and rafts of sea turtles. Sturgeon leapt from the water. The abundance of fish and marine mammals reported by Europeans in North American waters was once thought to be exaggerated, but paleoecological studies of Indian middens have shown the reports were probably true (skeletons in the middens indicated fish were abundant and large, though oyster shells at the bottoms of the middens were larger than those on top, and the fish caught shifted somewhat from predatory to herbivorous, showing that harvest by the native Indians had some effect). Later studies have shown that European riverine fisheries were as abundant as aboriginal North American ones up to about the year 1000, sea fisheries to 1300. The oyster beds of Chesapeake Bay were once capable of filtering the whole water column of the Bay every three to seven days. Chesapeake Bay averages 21 feet deep (20% of the bay’s bottom is less than six feet deep) and the water cleared of algae and other small plankton by the filter feeders (oysters, mussels, clams, worms) let in the sun and allowed the growth of sea grasses and rooted underwater algae. Several species of aquatic grass covered perhaps half the bay’s bottom, that is 500,000 to 600,000 acres, and helped oxygenate the bay during the summer months. Summer anoxia in the deeper parts of Chesapeake Bay is normal. It is caused by cold saline seawater pooling below the warm fresh waters from the rivers that flow into the bay. It was not a serious condition in the pre-contact bay. Since the bay is so shallow, winds and tides would break it up and move it around. However 15 million people now live in the watershed of the bay, and the equivalent of 4 million more are found in the 500 million broilers raised annually on the Delmarva Peninsula, which forms the bay’s eastern shore. The result is that the waters of the bay are tremendously overfertilised. (If they still existed, the seagrass meadows would have absorbed about half the nutrients in the sewage effluent now entering the bay.)

Spring along the east coast of North America is accompanied by an algal bloom. The vegetation of the tidal marshes has been ground up by ice shelves working back and forth over them; and recycled by feeding waterfowl. The rivers bring down fresh sediment and nutrients with the spring thaw. As temperatures rise, the organic material decays, feeding the growth of algae, which respond to the rising temperatures by dividing more quickly. The nutrients brought down by the rivers include silicon (from sand), needed by diatoms, the major component of a healthy bloom off these coasts, to form their skeletons. The growth of the algae may start as early as February and peaks sometime in June.

A working modern ecosystem (modern in a biological sense) is capable of turning the nutrients transformed into algae during the bloom into the tissue of higher plants and animals. Filter feeders like sponges, clams and oysters filter algae and bacteria out of the water; zooplanckton feed on the algae, small fish feed on the zooplanckton, larger fish eat the small fish, and so on. Beds of annelid worms, living in the bottom mud, their filter feeding parts protruding (nipped off by feeding fish, they regenerate) take in the living and dead material that filters down. Some of the nutrients produced by the bloom are used by the underwater grasses. Tides wash the bay’s water back and forth over tidal marshes, whose plants are now growing and also absorb nutrients. By the time summer anoxia sets in and makes the bay more vulnerable to oxygen depletion, the nutrients have been absorbed by the system. Any excess accumulates in the bottom muds, where they are recycled into the water column over the year.

As more and more forestland in the bay’s watershed was cleared for agriculture, and the bay’s wetlands were drained for agriculture and housing, the load of nutrients entering the bay rose and became continuous year-round, not a springtime pulse. Wheat-growing in Maryland and Virginia after the 1730s meant plows, draft animals, permanent pastures and fields, the spreading of manure, more land under cultivation. Eroding agricultural land sent more and more sediment into rivers. Sedimentation was at least 4 times pre-colonial levels from 1760 to 1860 and caused extensive shoaling in the upper parts of the estuary. (The Potomac River was 42 feet deep at Alexandria, Virginia, in 1794 and 18 feet deep—it would have been shallower except for dredging—in 1974.) Cutting the forest raised stream temperatures in summer and lowered summer water levels, both making life more difficult for anadromous fish (who in general need clear, cold water), and increased spring runoff by 25-30% over that of a forested watershed. In general, the higher the spring water flow down the Susquehanna, the greater the summer anoxia in the bay. More fresh water means more nutrients, more silt and more salinity stratification (warm fresh water sitting over cold salt water). Summer anoxia probably became a regular feature of the bay in the 1760s. The growth of a market economy for fish in the early 1800s meant stocks of herring and shad were overfished. Soon railroads and ice and canneries made it profitable to overfish oysters. Sail dredgers removed layer after layer from the reefs. Dredging went on day and night. Oyster fisheries in New State and Connecticut, nearer large markets, were in decline by 1808, when they began to import spat (young oysters) from the Chesapeake to replenish their beds. (In 1600 the estuary of the lower Hudson, including New York harbor, had 350 square miles of oyster beds and perhaps half the world’s oysters.) Chesapeake Bay produced an average of 15 million bushels of oysters a year throughout the 1800s, demonstrating the ecosystem’s tremendous resilience, which was thanks in part to the habitat developed over 4500 years of occupation by oysters. However production began to fall in the twentieth century with the physical removal of the reefs and oyster production is now 80,000 bushels a year.

Farming and dams also destroyed fish habitat. Dams held back sand, which was heavy and settled out, needed by the diatoms, the base of the spring bloom. They blocked fish runs, preventing anadromous fish like shad and river herring from reaching their spawning grounds. Sawdust and silt covered fish spawning beds. During the nineteenth century chemicals from tanneries, paper mills, steel mills and chemical industries, along with sewage and silt, polluted rivers. Sublethal toxicity from pesticides, herbicides, various hydrocarbons, oil, heavy metals (arsenic, copper, and lead were used as pesticides in the 1800s) affected the seagrassses and the fish and invertebrates, especially the animals’ larvae. Tides resuspended the pollutants and moved them up and down the bay. (Like spring high water, tidal movement was probably an advantage when all the water was moving was nutrients, in a nutrient-limited system.) Some oyster reefs were dredged for shipping channels. Removal of the oyster reefs removed places for oyster larvae to settle and also shelter for other invertebrates and fish. The vertical structure of the reefs had increased water turbidity over the oysters, reducing their exposure to low oxygen levels, enhancing their feeding, and preventing them from being smothered by silt. Overfishing eliminated the natural oyster fishery by early in the twentieth century. Artificial propagation kept the oyster fishery alive (at a much lower level) until a disease was introduced with asian oysters in the 1960s. With the oysters essentially gone, their ecological work of filtering the waters of the bay was also over and the growth of algae made the bay much more turbid. One of the remaining abundant forage fish, menhaden, a plankton eater, whose feeding might have helped control the algal blooms, was fished down for animal feed (those chickens of the eastern shore), to such an extent that when striped bass, a popular sport and culinary fish, recovered after closure of the bass fishery in the 1980s and 1990s, the growing bass starved, and died of opportunistc infections.

By the 1950s the systems that had maintained Chesapeake Bay had started to collapse. In the 1980s nitrogen entering the bay from its rivers amounted to about 220 pounds per acre of bottom (a heavy dose for farmland; several tens of pounds per acre also entered from the atmosphere). Phosphorus amounts (partly from detergents, partly carried with eroded sediments) were also substantial. By then the oyster fishery and the eel grass meadows had collapsed and many fish (shad, striped bass, river herring, yellow perch) and shellfish (including the oysters) were down to 5-10% of early levels. In 1833, 750 million river herring and 25 million shad were caught and salted in the bay; this number was reduced by 99% by 1878, mostly through overfishing; the fishery later recovered somewhat, but in general the condition of the rivers was too poor, their water quality too compromised, summer flow too low, and dams too many, to support fish. The bay’s food web started to shift from bottom dwelling organisms in a clear water environment (oysters, herring and their relatives, striped bass, eel grass) to a web dominated by microscopic algae, bacteria, and other single-celled organisms that live suspended in turbid water (a floating ecosystem), eaten by jellyfish. The normal way the ecosystem had dealt with the spring algal bloom no longer worked. The algae now grew until they ran out of nutrients, then died and sank to the bottom of the bay, where their decay used up the oxygen in the water’s lower levels (always in short supply in summer). The lack of oxygen smothered the grasses on the bay’s shallow shelves, already suffering from a lack of light from the algal turbidity and siltation; fixed animals like oysters and mussels also suffocated. Much of the primary productivity of the pre-settlement bay was channeled through the seagrass meadows, which removed nutrients, provided shelter and breeding habitat for fish and invertebrates, oxygenated sediments and the water, served as food for tens of millions of overwintering waterfowl (those flocks of canvasbacks, redheads, black ducks, shot commercially from 1870 to 1910, whose droppings helped fertilize the bay). When the meadows were gone, primary productivity shifted to algae. The reduced tidal marshlands absorbed less of the silt and nutrients in the water that washed with the tides back and forth across them. Eroding riverbanks sent down more silt. (Mudflats as a habitat increased.) More nutrients accumulated in the bay’s bottom muds, were recycled back into the water column, and fed more algae. The siltation and the increasing summer anoxia, along with commercial trawling, destroyed the bay’s benthos below about 22 feet (half the bay’s bottom) so what was a garden of sea pens, sea anemones, sponges, corals, clusters of worms, rooted algae and other bottom-dwelling organisms, that oxygenate sediments, recycle nutrients, and provide places to feed and hide for invertebrates and fish, became bare anoxic mud. Toxic pollution was also now a problem but nutrient pollution, overfishing and habitat destruction, which had been problems all along, were probably worse problems.

Changes like this are now universal in lakes and shallow seas around the world. One reason the Atlantic right whale population is failing to recover is suspected to be nitrogen pollution in Cape Cod Bay. The bay is a major winter feeding ground for the whales. (Sailing into the bay, the Pilgrims remarked on them.) It is thought microscopic plant growth in the bay overwhelms the zooplankton and other grazers on the plants (they can’t keep up) and the resulting mix of plankton is no longer rich enough to allow these whales (filter feeders like the oysters) to gain sufficient weight to reproduce: the whales have been turned into vegetarians. A similar explanation has been put forward to help explain the collapse of fur seals in the Bering Sea: as the Sea has warmed, fatty cold water fish have been replaced by leaner, warmer water fish; the new diet lacks sufficient calories for the animals (who, like the whales, spend much of their time in water near freezing). Overfishing may also be a problem in Cape Cod Bay: algal blooms can be enhanced by changes in the abundance of different species of fish; that is, by fishing pressure. Small plankton-eating fish preferentially feed on the larger zooplankton that eat (more of) the algae; if the numbers of the larger fishes that eat the small fish are reduced by fishing, the small fish increase, and eat more of the zooplankton, and it is easier for the algal population to get out of control. So we have a new estuarine world, perhaps harvestable for algal broth (whose overall productivity, boosted by fertile runoff, may be greater). Jellyfish also harvest algae (and, preferentially, the zooplankton that eat the algae), and some shrimpers off the coast of Georgia have begun harvesting jellyfish rather than shrimp, which are depleted there; there is a market for jellyfish in Southeast Asia. Seas whose biomass consists largely of algae and jellyfish (the Black Sea is an example) are what we would call destroyed ecosystems, though several hundred million years ago they would have been normal. Of course the usual modern changes also occurred in Chesapeake Bay’s watershed. Most rivers along the east coast of North America carry substantial amounts of oil (poured down drains, washed off roads and parking lots). Polycyclic aromatic hydrocarbons in the oil and in the fallout from power stations and automobile combustion cause deformities in developing fish embryos that kill many of them before they reach adulthood: a tax levied by industrialization. (Birds pay a similar tax levied by lighted windows, cats and cell phone towers, amphibians one from agricultural chemicals in surface waters.) Dams cut down on the supply of silicon needed by diatoms, a major part of the original algal system, to build their shells. This together with the excess nitrogen has shifted the population of planktonic algae toward the dino-flagellates, some species of which are poisonous and form the so-called red tides. Riverine sand and silt, like nitrogen and phosphorus, are both nutrients and pollutants; it’s the timing and the amount that matter.

* * *

Farming is a complete break with the earlier landscape. Most native plants and animals are eliminated. (As a rule of thumb, modern settlement reduces native flowering plants, birds, amphibians and mammals by 90-95%). Water runoff increases: flood levels rise, but summer low flows are lower. The water table falls. Soil and nutrients run off with the water and also blow away. Wind speeds near the ground are higher. In temperate regions the ground is left bare for seven or more months a year. On bare ground the snow melts more quickly. The ground also dries out and warms up faster in the spring (a help to farmers). Soil microbial activity and nutrient turn-over and loss are higher in the warmer soils. Loss of soil nutrients by rain increases. Cultivated soils also lose carbon, which adds to the carbon dioxide to the atmosphere. After a few decades of agriculture in temperate regions, soil organic matter falls to half its level in unfarmed soils. Levels of soil organic matter fall more quickly in the tropics. This makes the soil more erodable. When bare ground is covered with snow, its reflectivity is high, compared with forestland, so the ground gets colder and the frost goes deeper. Where I live water pipes should be put four feet below open ground to avoid freezing, deeper if the ground is compacted (as under a road). But frost in the woods is rarely a foot deep, probably partly because the winter water table in the woods is so high, partly because the trees shelter the ground from the wind and from radiant losses on clear cold nights.

The ecological effect of agriculture on watercourses, or on the plants and animals of a landscape depends on its extent. Limited fields, in a larger natural landscape, like those of the Native Americans or of the early period of English settlement in New England, create a more varied environment, with more edge habitat, and more game animals, which are for the most part creatures of the edge. The typical eighteenth century subsistence farm, with its weedy shrubby fencerows, different crops in different small fields, uncultivated woodlots, marshes and steep or rocky areas, provided habitat diversity. Much of the burning of Native American peoples was intended to create the berries and browse of early successional landscapes that game animals favor. Forests of nut-bearing trees were created by fire, and people harvested the nuts, and the animals that ate them. Slash-and-burn agriculturalists in tropical forests over time modify much of their forestland whose soils are suitable for agriculture; but slowly, with long rotation times between clearing a plot again. After a few years crop yields fall dramatically as the nutrients in the soil and in the ash from the burned vegetation are used up. (In Indonesia rice yields fall 80% between the first and second crop.) Native plants (weeds, here) take over and the grain, tubers, and vegetables in a garden are replaced by planted food trees. Those are replaced 20 to 40 years later by forest trees; so much of the forest seems like primary forest to the untrained eye. Much of it would be, since not all soils were considered cultivable. But in the Amazon, for instance, much more forestland was probably cultivated forest (including planted trees) than first thought; man-made terra prieta soils may occupy 10% of the landscape and forests with a high proportion of useful trees, many planted, another 20%. The Iroquois of central New York State practised a temperate variety of swidden agriculture. With perhaps 10,000 acres under cultivation yearly, they might have cultivated a considerable percentage of the light upland soils they considered usable over the thousand or so years of their occupation (the fields also had to be near good village sites and sufficient supplies of firewood); but because of the time involved, their overall effect on the larger environment was small. (Local effects may have been considerable.) Fish, whose abundance is a sign of healthy waterways, provided a major part of their food. Slash-and-burn systems, like any agricultural system, break down when the human population gets too high. Fifty people per square kilometer—which seems a very high number—has been called the limit in Indonesia. Since the population of the Iroquois was small (12,000 to 20,000 people on several million acres), their overall effect was little different from the effects of the animals and plants which whom they shared the landscape. The feeding habits of elephants and moose also alter landscapes; so do earthmovers like prairie dogs. Elephants will destroy acacia forests in Africa and leave behind them grassland.

In such earlier worlds, with much smaller human populations, the influence of people on landscapes was more like that of other beings. In general, plants control the distribution of resources across a landscape; nutrients accumulate under shrubs in deserts, making life difficult for grasses growing in between; most of the nutrients in tropical forests are found in the vegetation. Plants control fire cycles (some encourage them, some discourage them) and the height of water tables; in Australia, eucalyptus trees, by taking up rainfall, prevent salt from deposits lower in the soil from migrating upward and make life in the upper levels of the soil more favorable, and surface runoff into rivers less salty. In the Northeastern United States nestholes dug out by yellow-bellied sapsuckers are made possible by fungi feeding on aspen trees. The nestholes are subsequently used as winter shelter by chickadees, nuthatches and white-footed mice. Beavers turn small streams into a series of shallow ponds, transforming the watery habitat into a paradise for freshwater invertebrates, fish and amphibians. The ponds neutralize acid surface waters, trap sediment and agricultural nutrients, and improve the habitat for songbirds and wolves. Deer and meadow voles influence forest succession by browsing some tree seedlings preferentially; the seed-eating white-footed mice prefer the large seeds of oak and pine (but probably don’t influence those trees’ abundance). Squirrels may affect the evolution of nut trees by selecting, burying and abandoning the largest, thinnest-shelled nuts. Kangaroo rats in the Chihuahua Desert of Arizona suppress large-seeded grasses by eating their seeds; their feeding helps maintain the shrub-dominated desert. Plants set the scene: plants change levels of atmospheric gases; modify water and nutrient cycles; engage in chemical warfare; promote forest fires; provide shade; and alter wind speeds, relative humidity, the penetration of frost and a landscape’s albedo.

Agricultural people manipulate the landscape by changing the plant cover, by hunting grazing animals (who also affect plant cover), and by introducing grazing animals of their own. Where agriculture becomes extensive, the natural habitat is reduced and simplified, and agriculture becomes the dominant influence on the environment. Erosion rates on cultivated crops are 10 to 100 times those in forests. Numbers vary. A study of Devonshire clay loams found that heavily grazed permanent pasture would absorb 0.1 millimeter of rain an hour (very little: a quarter inch a day), bare ground 4 millimeters, freshly seeded ground 11 millimeters (a heavy rain, about 0.4 inch an hour), freshly plowed, uncompacted soil 50 millimeters, and woodland soils 180 millimeters an hour—this is a flood, 7 inches an hour; the figure is hard to believe. The pasture must have been very compacted because generally land in pasture or cover crops absorbs 2-4 times the rainfall of land in cereal grains. Measured erosion rates from the United States range from about 60 pounds of soil per acre per year for undisturbed forest (that is, almost nothing, 2 bucketsfull), to 600 pounds per acre per year for grassland (still extremely low), 7 tons (14,000 pounds) per acre for cropland, about 12.5 tons per acre for logged forest (compared with 60 pounds per acre for unlogged), to hundreds of tons per acre for an active surface mine or a construction site. The maximum soil losses compatible with sustainable agriculture (where soil loss equals the rate of soil formation) are site specific, and vary from 0.4-0.8 tons per acre to 8 tons per acre; but generally losses greater than 4 tons per acre result in a loss of topsoil and thus also of fertility (some writers put this number higher, up to 10 tons per acre, or one large dump-truck load of dirt). Four-tenths of a ton per acre (one ton per hectare) means loss of an inch of soil every 250 years, while an erosion rate of 4 tons per acre means loss of an inch of soil every 30 years, on common agricultural soils. The yearly cost to the farmer of a loss of 4 tons of topsoil per acre is small (40-50 pounds of nitrogen, 10 pounds of phosphorus, plus a small loss of soil rooting depth and water holding capacity). The farmer’s cost is probably much smaller than the cost to the municipality or utility in whose reservoir the soil ends up.

Erosion rates vary with rainfall and land use. Topsoil is maintained by a balance between erosion and the disintegration of the underlying rock. A soil’s nutrients come from weathering of the rock, the decay of plant and animal material, and biological fixation. Rocks are dissolved (“weathered”) by the carbonic acid in rainfall; and by the acids secreted by plant roots, soil fungi and microorganisms. They are broken up by frosts. Small bits of rock are reduced further by the acid in the guts of worms and by the grinding action of earthworm gizzards. As Darwin pointed out, where worms are abundant they shape landscapes, raising the soil level, burying surface features, and moving soil downhill. The topsoil in abandoned English fields consists of worm casts and rock fragments. If more soil is produced than erodes, the topsoil thickens and eventually buries the underlying rock beneath the reach of active soil-forming processes. Then topsoil formation stops. So the thickness of a soil is characteristic of a certain landscape (its underlying rocks, its slope, rainfall, plants, animals, microorganisms). Some soils are deposited by wind, such as the loess soils of the north temperate regions, formed during the last glaciation, which created the great grain-growing regions of North America and Eurasia. So soils, landscape, and plant cover evolve together, often with characteristic animals and fish. A soil’s living organisms affect its structure, color, chemistry, the way water flows through it. The phosphorus and sulfur in soils comes from weathering of rocks, the nitrogen from biological fixation of gaseous nitrogen in the atmosphere. The slow accumulation of organic matter in soils binds them together and makes them harder to erode. Rocks rebound as the weight over them lessens (so hills remain much the same height as their soil erodes) but erosion slowly shapes landscapes. Once a landscape has become heavily agricultural, the continuing effect of agriculture on it depends on what crops are grown, how crops are cultivated, and where natural land is allowed to remain.

From the 1850s to the 1990s about 2.5 billion acres of forests, grasslands, and wetlands were converted to farmland worldwide, with tremendous effects on natural habitats and watercourses and also effects on global climate. The land conversion produced 30% of the anthropogenic carbon dioxide added to the atmosphere during that time. Currently agriculture contributes about 25% of the gases causing global warming. These include nitrous oxide from fertiliser, methane from rice paddies and the digestive tracts of cattle, sheep and goats, carbon dioxide from cultivated soils. While little new land has been cleared in the United States since the 1950s, cultivated American landscapes have been more and more cleaned up. Since colonial times, every increase in cash or animal power on the farm has meant more changes could be made in the original landscape: wet spots in fields, or marshes drained; woodlots (no longer needed for timber or fuel) cultivated. On traditional subsistence farms, weedy, shrubby field edges, different crops in small fields, uncultivated woodlots, shallow ponds, steep and rocky places provided habitat diversity. With the agricultural prosperity that followed the Second World War, the agricultural landscape became much less diverse. Home orchards and woodlots were converted to cropland, hedgerows between fields removed. Perhaps half of the agricultureal landscape in the Middle West was in sod or sod crops from the 1860s to the 1950s; as tractors replaced horses in the 1950s, soybeans replaced sod crops and pasture. (The loss of woodland was hard on birds. Many year-round resident birds of central Iowa shun blocks of habitat less than 35 acres and need 150 acres for successful breeding. Neotropical migrants that breed in Iowa need several thousand contiguous acres if the area is to be capable of producing a surplus of birds and not be just a sink for surplus populations from elsewhere. In woodlots under 500 acres in Iowa exterpation and recolonization by migrating songbirds is common; over 75% of the ovenbirds in woodlots less than 360 acres are unmated. So the birds come and go.) Along those hedgerows, fiberglass and steel have replaced the rotting wooden fence posts, in which red-headed woodpeckers once hollowed out nest holes. The holes were occupied the next season by bluebirds, who raised their broods on leaf-eating worms scavenged from the corn. Pollinators no longer fly out to pollinate seed crops, leaf-eating caterpillars no longer feed on weeds, predatory insects no longer capture aphids, parastic wasps no longer lay their eggs in the bodies of various beetles and caterpillars. The average square mile of farmland in Ohio and Illinois in the 1940s contained over 10 miles of hedgerow habitat (most of Ohio and about 40% of Illinois was forested when settled). Hedgerow trees were animal and wind planted. Their trees included black cherry, elm, hickory, oaks and honey locust. Chokecherry bushes and native shrubby dogwoods lapped up against the trees. The hedgerows slowed winds and sheltered birds, small mammals, and pollinating and predatory insects; they protected small drainages. As crop prices stayed low or fell, while the costs of farming rose, small farmers failed and farms became larger. Most crops need attention at the same time of year and larger fields were faster to plow and cultivate. Small in-field wetlands were drained or filled, removing more habitat and also the ecosystem services of aquifer recharge and nutrient removal. Streams were re-routed to square off fields, their bank-holding vegetation removed, and turned into silty drains. The silt ended up in downstream waterways. (Much of Iowa drains through channelized streams and ditches.)

As agriculture became more commercial, cultivation became more intensive. Crop rotations were reduced or eliminated as fertilisers and pesticides let farms specialize in row crops. (That it takes the manure of 1 cow to fertilize 2 acres put a certain limit on farm size; it takes the manure of 80 cattle to farm a quarter section, 160 acres. Farmers with the machinery to cultivate more land needed more cattle or synthetic fertiliser.) As farming became non-organic, farm animals went elsewhere. From the prairies of the Corn Belt, cattle moved west to feedlots on the plains and were fed irrigated corn; chickens went to the Southeast to live in huge chicken houses, some hogs stayed in the Middle West, but were now raised in confinement. In the 1990s, 75% of Nebraska cropland was in continuous corn, much of it irrigated with water pumped from the Ogallala aquifer. Thirty percent of cropland in Iowa and Illinois is in continuous corn, 60% in two years of corn, followed by one year of soybeans. Much of the farmland in both states is drained. Drainage greatly improves farmland, resulting in better soil aeration, faster decomposition of organic matter, deeper rooting of crops, a warmer soil. It also increases rates of water and nitrogen runoff by several times. (The nitrogen comes both from applied manure and fertiliser and from the stored nitrogen in the soil; losses of stored nutrients in drained soils take decades to deplete.) Both corn and soybeans are very erodable row crops; and corn rotated with soybeans can result in much higher rates of erosion than continuous corn. Such cropping is only possible with large applications of fertilisers and pesticides. (In row cropping the plants are seeded in straight rows, the rows far enough apart so the growing plants can be cultivated with a tractor.) Fertiliser use was given a boost by the Second World War. During the war hydropower plants along the Tennessee River were used to manufacture ammonia for explosives; after the war the ammonia production (now surplus) was sold as fertiliser. (The demand for it now insured, ammonia is currently synthesized from natural gas and shipped as a liquid by pipeline from plants in Texas, Oklahoma and California.) Factories that had been making tanks were retooled for new markets in cars and farm tractors.

Modern commercial agriculture depends on applied chemistry, applied genetics and machine power: the use of artificial fertilisers, pesticides, irrigation, and crops bred to produce large yields with large inputs of water and nutrients. So-called organic or regenerative agriculture depends on applied biology. In any agriculture, but especially in an organic agriculture, soil microbes (the mostly unknown archaea and bacteria) and soil fungi are the basis of productivity. A small part of soil organic matter, soil microbes hold its largest pool of labile nutrients. Many crop plants have connections with mycorrhizal fungi, which supply them with water and micronutrients, in return for sugars. This connection is suppressed by the use of nitrogen fertiliser, which makes the plants withhold their secretions from the fungal hyphae. Thus artificially fertilized plants are more suscepible to micronutrient deficiencies and to drought. Many crop plants also have associations with nitrogen-fixing bacteria, which supply them with nitrogen in return for root secretions. (Breeding could increase this number.) In an unfertilised soil the nutrients contained in dying soil microbes feed plants. The microbes also recycle nutrients in dead and dying crop plants and make them available to the new crop.

The soil ecosystem as a whole (fertilised or not) depends on microbes and organic matter, its source of nutrients. Less organic matter in a soil means less activity among the soil biota, less nutrient cycling, and less vigorous growth of crop plants. Organic matter helps soils retain moisture, improves their tilth (polysaccharides secreted from soil fungi glue soil particles into clumps); the chemistry of soil organic matter helps liberate nutrient particles from clays. Both crop rotations and manure raise soil organic matter. Manure also raises nitrogen levels in soils (much of its nitrogen is temporarily immobilized by soil microbes), while synthetic nitrogen fertiliser tends to be quickly lost from the soil, approximately half of it used by crop plants, the rest lost to surface runoff, ground water and the atmosphere. A 15 year study of corn and soybean cropping at the Rodale Institute in Pennsylvania found no significant difference in crop yields among plots fertilised by rotation into legume sods, fertilised with manure, or with synthetic fertiliser. The soil carbon content of the manured plots increased three times, and of the plots rotated into legumes five times, over the carbon levels in the conventionally farmed plots. Other studies have found organic agriculture producing yields comparable to conventional agriculture in normal years, but 30% higher in dry years, and greatly reduced levels of erosion in plots where the farming practice included rotation into legume sods. In two wheat farms in the Palouse region of Washington State, one organic since 1909, one plowed in 1908, but farmed conventionally since 1948, wheat yields were similar. The organic farm grew a green manure crop (usually alfalfa) every third year. The conventional farm lost six inches of topsoil from 1948 to 1985. The organic farm retained its topsoil, which had twice as much organic matter, more moisture holding capacity, more available nitrogen and potassium and a greater density of microbes than the conventional farm. Of course, because the conventional farm didn’t rotate its fields, it produced more total wheat over the period, but at the cost of its long-term viability.

Corn and soybeans replaced the mixed farm with its cattle and hogs. Both corn and soybeans are very erodable row crops; soybeans fix nitrogen but only enough for the crop itself. The traditional four-field crop rotation of the American Middle West consisted of legume hay, corn, oats and winter wheat. Approximately 25% of the land at any one time was in sod (including small grains as sods brings the number to over 50%). The animals (including the work horses) were fed the hay, and some of the grain; some grain was sold. Feeding grain (converting it into animal flesh) was a way (through an additional investment of labor) of increasing its value. The corn and oats were manured; and livestock grazed the new growth of the winter wheat in the fall and the stubble of the harvested grains. With the animals gone and the land in row crops, erosion and soil degradation have increased, with negative effects on rivers, lakes and connected ocean waters. Soil fertility declines under synthetic fertilisers unless straw or manure (some sort of carbonaceous material) is added to the soil. However, keeping animals and rotating crops into legume hays reduces the amount of land in the (easily marketable) cash crop. Cropping systems become more complex; more farm equipment may be required. Marketing other crops or animals may be difficult. Cultivation of corn and soybeans was intensified in an attempt to increase profits; a capitalist farmer and his banker try to maximize the returns on capital investment: in this case, land. (But until recently—2008—commodity crops did not return much per acre.)

More intensive cultivation has many effects. Hayfields, which once provided breeding grounds for grassland birds, are now mowed too frequently for the birds to finish a reproductive cycle; meadowlarks, bobolinks, upland plovers, savannah sparrows cannot use modern hayfields. Partly as a result, grassland birds are among the most endangered in the United States today. Pollinators such as bees, moths and butterflies (already deprived of hedgerow habitat) cannot take advantage of a flowering crop: most recommendations are for cutting at one percent bloom. Isolated patches of prairie orchids are pollinated by moths that must cross miles of chemically treated fields to reach them, so many go unpollinated. The situation could be improved by timing mowings so as to give the birds a sufficient period to breed; this means a lower forage quality in the first cut, which in my area would have to be harvested about two weeks later. (I’m not sure what the cost in lost quality would amount to, perhaps $5-$10 per acre.) Some nearby habitat, such as the grass waterways meant to capture water and nutrients running off the field would have to be left unmowed until mid-summer to provide cover for the young birds. Since birds return to specific nesting sites each year tricks like this might work; though overflow populations that tried to breed in adjoining, conventionally managed fields would be unsuccessful. One of the better innovative farming schemes, so-called rotational grazing, lets animals rather than people harvest the grass. Grazing in this case more closely mimics what happens on a natural grassland. In rotational grazing schemes less hay is cut for the animals, few if any row crops are grown, and more land goes into pasture. The animals are kept on pasture for as long as it is available (7-12 months, depending on the climate). Each pasture is grazed intensively for several days and then left for approximately a month to regrow. Birds use the pastures, which also have a much more varied invertebrate and plant life, and probably function better as whole ecosystems. (Function may partly depend on the plants and animals: the structure and water-holding capacity of the natural prairie established within the circle of the Fermilab Nuclear Accelerator near Chicago is dramatically better than that of nearby planted pasturelands. Cattle graze the pasturelands, buffalo the prairie.) Rotational grazing schemes produce meat and milk from grass (which cattle are adapted to eat) rather than from corn and soybeans (which in large quantities make cattle sick; feeding cattle corn is like feeding people on a diet of Coke and Milky Ways.) The animals’ manure and urine fertilise the pasture and add to their organic matter; the grazing raises the productivity of the grasses. In some schemes, chickens are brought in a few days after the cows leave. The chickens eat insects and the remaining grass and scratch apart the cow patties to get at the hatching fly larvae. So the chickens spread the manure and keep the fly population down and the farmer harvests eggs and chicken flesh from his pasture. Joel Salatin’s Virginia farm produces 35,000 dozen eggs, 10,000 broiler hens, 40,000 pounds of beef, 30,000 pounds of pork, 1200 turkeys and 1000 rabbits on 100 acres of grass and 450 acres of forest and a small amount of purchased corn. The farm was eroding and wornout when Salatin’s father bought it. Now erosion from the grassland is essentially zero. (Loss of nitrogen from pastureland is approximately 35 times less than that from fertilised row crops.) In a nice twist, Salatin periodically scatters corn over the surface of the manure that builds up in his cow barn over the winter. In the spring pigs are let root in the manure for the fermenting grain. The pigs have a good time, and in doing so turn the manure and speed up the composting process.

Organic farmers are interested in keeping nutrients, including carbon, in their soils. Some writers claim organic agriculture could absorb 8-17% of U.S. carbon emissions. Farms that don’t rotate crops must use chemicals. Nitrogen fertilisers raised farm output substantially when they came into common use. Much of the doubling of the world’s output of food from 1950 to 1980 is thought to have come from the application of artificial fertilisers, but perhaps half came from increased irrigation, some came from the ability to grow continuous crops of grain on already existing farmland, and some from the bringing of another billion aces of land into cultivation from 1920 to 1980, mostly in South America. Plant breeding also raised output. American corn yields rose 72% from 1975 to 2005, mostly thanks to the development of higher-yielding varieties, many of which needed large amounts of nitrogen. The American corn crop has grown steadily at 2% a year for some time. Nitrogen fertilisation raises the nitrogen content of leaves and makes them more attractive to leaf-eating insects, which are more vigorous and fecund on their rich diet. Fertilisation also seems to lower the production of phenolic compounds in the leaves. The phenols act as insect repellants and, paradoxically, as anti-oxidants, beneficial to us, when we eat the plants.

Growing the same crop in the same field year after year lets populations of pests and weeds build up, sometimes catastrophically. In the Red River Valley of North Dakota, growing continuous crops of wheat during the twentieth century resulted in outbreaks of fungus disease so uncontrollable that wheat had be be abandoned as a crop. Continuous grain requires pesticides and herbicides. In the 1990s American farmers used almost a billion pounds of pesticides on their crops. Farm chemicals have effects on farmers (farm families have among the highest rates of cancer in the United States) and on wild animals; they kill non-target insects and invertebrates, including pollinators, and insects used as food by birds and other animals. Increased use of agricultural chemicals including dieldrin, DDT, malathion and atrazine have reduced populations of the northern leopard frog in the United States by 90%. Dieldrin and DDT are now banned in the United States and Canada, but still manufactured and exported, and still arriving on the air from Africa, Asia and South America: freshwater bogs in the eastern United States are still accumulating DDT. The chlorinated hydrocarbons reduce the frogs’ immune systems almost to zero; thus they are easily parasitized; this results in the large population of deformed animals. Atrazine also interferes with the frogs’ sexual development.

The tidying up of the countryside (plowing to field edges, spraying field margins, draining wet pastures) that machine power and farm chemicals made possible eliminated much of the habitat of the harvest mouse in Britain between 1970 and 1990; the animal’s food plants and insect prey were gone. The pollinating insects and the insect predators and parasitoids of crop pests require a complex architecture of hedgerows, wildlands, crop mixes and weedy field edges to survive in useful numbers. They need food and habitat apart from that provided by the crop plants. Evergreen blackberry plants at the edge of California vinyards provide shelter and prey for a wasp parasite of vine leafhoppers sufficient to keep their numbers high enough so that when winter is over and the leafhoppers once again become a problem in the vinyard, there are enough wasps to make a difference. In fruit orchards of the Pacific Northwest, strawberry plants and wild roses provide winter habitat for wasp parasites of leaf rollers. A few Canada thistles in the edge of an Iowa field provide places for the diseases that attack its leaves and roots to overwinter; and a place for the painted lady butterfly (whose larvae eat thistle blossoms) to breed. A field without insect pests is also a field without their predators, a situation knowledgeable farmers find alarming. One needs both pests and predators. It’s a question of relative numbers: of competitive adjustment.

* * *

In the Middle West, river floodplains and wet upland prairies were settled last; the forested lands on the slopes were thought to have the best soils. In an ideal world, much of the floodplains would never have been settled. While they are usually good agricultural soils, they are also places where nitrogen is removed by denitrifying bacteria from surface or ground water. Carbon-rich, damp sites at the junction of aerobic and anaerobic conditions are major sites for denitrification. Such sites are usually found along the riverbank but may be located all over a floodplain, and throughout upland fields. Floodplains also remove topsoil and nutrients from water flowing over them (so-called “overland flow”); floodplain forests shade and cool edges of streams (and the groundwater that flows beneath them into streams). Sufficient forestland along streams will convert much of the overland flow into subsurface flow and remove most of the nutrients in the water; so the water entering the stream will be cleaner and cooler and its movement into the stream not so rapid (that is, more like the original stream conditions). Estimates of the width of the strip required to perform such services vary from 60 to 200 feet; and the strip should contain any ponds or wet depressions connected to the river. Studies in Tennessee show that 6% of an agricultural watershed in strategically planted forest will convert half of the agricultural run-off into subsurface flow and remove a considerable proportion of the excess nutrients. Studies in North Dakota show that 15% of the landscape in natural land will support a residual population of wildlife sufficient to allow hunting and thus bring in additional income. Such lands also support insect and bird life helpful to the farmer: Eugene Odum recommended 40% of any landscape be left undisturbed; some organic vinyards in California leave half their land as wild land. Unplowed grassland along streambanks also helps slow water flow and remove nutrients; so do grass waterways that carry surface flow from fields; both can be mowed or lightly grazed and maintain their value as nutrient traps. All the same, some nutrient pollution from farmland is unavoidable. Heavy rains (especially in winter, when the ground is frozen) tend to channelize the flow of water, reducing its contact with fallen branches, leaf litter, logs and grass stems, and thus send more run-off, with its nitrogen and phosphorus, directly into streams. Some of the nutrients and water may be stored by downstream wetlands connected to the river—another reason to preserve large parts of the natural river. River bottomlands are important habitat for wildlife and also serve as corridors for animals and plants moving among different habitats. The species diversity of plants is greater in lands connected by corridors, probably because animals that spread seeds can move more easily among them.

Such numbers give us a means of designing new landscapes. Restoring semi-natural landscapes along streams is probably essential for stream health. Such landscapes are also useful for the farmer (as nutrient traps, sources of pollinating and predatory insects, habitat for birds and animals that control plant-eating insects and rodents). They are not extensive enough to restore the plants and animals of the original environment; but this is not an argument against them: like current farmland, these are new landscapes. The neotropical birds that once nested along Northeastern and Middle Western rivers need thousands of acres of unbroken forest for successful nesting, partly because of nest predators that prefer edge environments (skunks, feral cats, opposums, raccoons, crows, jays, all more abundant because of the absence of large predators), partly because the forest near an edge is drier. One study found that the mass of soil insects and invertebrates in the interior of large blocks of woodland in ovenbird habitat in southern Ontario was 10 to 36 times greater than that near the woodlands’ edge. So apart from losses caused by nest predators and nest parasites like the brown-headed cowbird (also birds of fragmented habitat), ovenbirds find better foraging and raise more young in large blocks of forest. Since the effects of the edge (changes in air temperature and humidity, in ground moisure levels) penetrate from 100 to 200 yards into the woods (some scientists now say up to 400 yards, which makes all but enormous blocks of woodland edge environments, but the effect runs along a continuum), bands of woodland 50 to 200 feet wide along a stream will not provide good reproductive habitat for many neotropical migrants (the most common and abundant birds of the eastern forest), and thus the habitat will lose the ecosystem services they, if abundant, would provide. (A full compliment of birds in the eastern deciduous forest, by helping to control grazing insects, is calculated to be worth about $7 an acre, or $5,000 per square mile of forest per year.) Self-maintaining populations of large animals like elk, wolves and mountain lions would probably need more habitat than protected riverine areas would provide. (While Eurasian wolves survive in rural Tuscany and mountain lions on the outskirts of Berkeley, California, North American wolves in lightly settled parts of Canada move out when the human population rises above three per square mile.) But protected stream corridors would provide habitat for many songbirds, small mammals, amphibians, predatory and pollinating insects, as well as other invertebrates and some larger mammals and predatory birds. Owls (nesting in owl boxes) and raptors (nesting in tall forest trees) might become a natural part of the farm scene (preying, say, on English sparrows, starlings, rats and feral pigeons; in a related matter, California organic vinyards use falcons to scare away starlings, golden retrievers to sniff out female vine mealybugs, and chickens in mobile coops to control cutworms and weeds). The purpose of the strip is to control the temperature and flow of water into the stream, remove excess nutrients, and to provide habitat for animals and insects useful to the farmer; in combination with other practices, riverside wildlands may provide income from hunters; they also give the landscape connectivity. For all these purposes, the wider the strip, the better. The nutrient-removing function of riverine lands is largely accomplished by bacteria and depends on soil temperature, soil carbon levels, and the degree of aeration, as well as on uptake by plants; strips should be calculated so as to capture most of the nutrients flowing off a farm. Riverside wetlands are capable of removing 90% of the nitrogen and 80% of the phosphorus in groundwaters flowing through them. (Most of the phosphorus is not dissolved in groundwater but attached to fine clay particles and must be filtered out of overland flow by the surface of the ground.) In an ideal farm landscape, all lands functioning as nutrient traps would not be along streams; some would occupy usable farmland. Some farm crops, like hazelnuts grown for oil in Minnesota, can also function as nutrient traps; so can tree crops, such as pines or poplars grown for pulpwood, though the overall ecological value of such monocultures is probably less.

Farming systems matter in stream health. Natural streamside filter strips can be overwhelmed by large amounts of nutrients (especially phosphorus), and even if they remove 90% of the nitrogen that moves through them, 10% of a very large number is also a large number. So controlling the nutrient levels in rivers and lakes also means modifying agricultural processes. Erosion under cultivated annual crops (corn, soybeans, lettuce) on gently sloping land (slopes of less than 5%) may be 10 times that in forests, several times that in grassland; on steeper slopes, erosion under row crops is several hundred times that under forests. Erosion of soil, and of the nutrients and pesticides carried off with it, is a concern for rivers and lakes; for water supply reservoirs, which must supply high-quality drinking water and whose lifetimes are measured in the amount of silt that enters them; for estuaries, such as Chesapeake Bay and San Francisco Bay, whose marshes, underwater meadows, and beds of worms and molluscs are susceptible to siltation, and whose waters now receive several times the nitrogen of an average fertilized farm field. Erosion is also a concern for the sustainability of agriculture at a site. Fertility declines about 6% with each inch of topsoil. Topsoil is being lost by American agriculture at about 6 inches per century. Iowa loses 0.5 inch a year. In Georgia and Tennessee losing 6 inches of topsoil reduced crop yields by 50% (somewhat more than expected). A 40% reduction in the nitrogen now carried by the Mississippi is necessary to begin shrinking the dead zone in the Gulf of Mexico. The cost of reservoir siltation alone exceeds the cost of agricultural practices that reduce erosion. But farmers don’t pay that cost, or—except on Conservation reserve land—receive compensation for using practices that reduce siltation. Generally in the United States, a dollar invested in soil conservation yields $3 in increased crop yields and saves $5 to $10 in the external costs of farmland erosion, such as dredging rivers, building levees, replacing dams, and other flood control works.

Erosion-limiting agricultural practices include crop rotation; terracing and contour planting; strip cropping (strips of row crops alternating with strips of sod crops); minimum tillage (seeding into the stubble of last year’s crop); stubble mulching (not plowing fields in the fall, but leaving the stubble, preferably with a cover crop seeded into it, until spring); restoration of hedgerows (which break the wind); plowing of grain straw back into the soil (incorporation of straw in temperate climates leads to a 40-50% increase in the soil biomass of carbon and nitrogen over 20 years and much improved soil tilth). In dryland farming, the appropriate degree of tillage helps control erosion, maintain soil porosity (and thus the ability of water to infiltrate) and increase the root density of the crop plants. Soil conservation trials in Texas, Missouri and Illinois slowed soil loss by 2 to 1000 (!) times and increased yields of cotton, corn, soybeans and wheat by up to 25%. The costs of soil conservation are paid back by increased crop yields, while over the long term erosion costs the farmer money. A small soil loss, 6 tons per acre, means a loss of 110 to 130 pounds of nitrogen and 11 to 18 pounds of phosphorus per year, and the loss of an inch of topsoil in 20 to 25 years. These nutrients will have to be supplied by the farmer. One-half of the fertiliser used in the United States replaces nutrients lost by previous erosion. Erosion over a site-specific amount also results in a loss in the soil’s rooting depth, in its water holding capacity, its degree of aeration. Erosion of farmland soils is now 10 to 40 times the rate of soil formation. Such rates of erosion have meant the end of former civilisations. It seems reasonable to think that agricultural activity should not generate soil movement beyond that for which cultivation of the land can compensate; that is, soil fertility is a matter of public interest.

The movement of soil downhill is ineluctable; what matters is its speed of motion. Soils exist in a dynamic balance between depletion (through leaching, erosion, gaseous losses, removal of nutrients by plants and animals) and renewal (through weathering of rock, decomposition, nitrogen fixation, inputs from the atmosphere and from plants and animals). Natural soils vary in their success at maintaining this balance. (Thus soils are thicker or thinner.) Farmland topsoil is re-created by incorporating crop residues and manures, by catching the dust that falls out of the air, by adding rock powders (ground limestone, rock phosphate), by plowing in sod crops, by slowly incorporating the upper layer of the subsoil. The sediment, phosphorus and nitrogen that enters streams in hilly prairie landscapes can be reduced by 80% by crop rotation that incorporates hay crops and small grains (wheat, oats and barley function as annual sods, though with runoff rates up to 8 times that of a prairie); by adding 300 feet of grassland buffer along streams (not a minor amount, but the land can be grazed or used for a late hay crop); and by feeding animals on pasture rather than in feedlots. Legume hays not only reduce erosion; they also restore soil tilth and water-holding capacity and increase the soil’s pool of available nutrients. Wheat and corn rotated with clover hays yield 30-50% more per acre. The clover, or clover-grass mixture, is pastured, or harvested as hay. On flatter prairie landscapes, the same practices also reduce nutrient runoff and the buffer strip along the streams can be narrower.

Rotating land into legume sods reduces the land in the cash crop (usually soybeans or corn). Hay crops bring animals back to the farm. Corn and soybeans are for the most part grown to feed animals. Something like 65% of the American corn crop goes for animal feed, while something like 40% of grains and soybeans grown worldwide go to feed animals. Another way to use hilly prairie landscapes is to raise cattle (beef cattle or dairy cattle) on grass, by rotational grazing. Rotational grazing divides up the land into paddocks, which are grazed for short periods (three days to a week), and then rested (for perhaps a month); the manure and urine from the animals provides fertilisation, and the resting period lets the grasses regrow; each paddock requires a supply of water for the animals. Paddocks are periodically rested for longer periods. Then they are not grazed until the end of July; this lets them reseed, the roots of the grasses develop, and results in better long-term productivity. It also lets some native bunch grasses survive and makes better habitat for grassland birds. Surface water ponds in the prairie pothole region of Iowa, Minnesota, and South Dakota once provided 30 million acre-feet of water storage per year, water that now flows through drains into streams and rivers feeding the Mississippi. The region was a major North American habitat for breeding waterfowl, a “duck factory.” Much of the region has been drained. In this region, blocking the water flow at the bottom of a depression (blocking the drain) provides a pool of water for the cattle. Paddocks cluster around the restored pothole. The wetland edge is then used by ducks, grassland birds, and cattle (which tend to graze around the duck nests). Pintail ducks and geese seem to prefer the grazed vegetation (perhaps as food, perhaps because they can see predators at a distance). This mimics the natural situation: elk and buffalo once grazed these landscapes. Rotational grazing becomes a functional replacement for corn and soybeans; one ends up with the same crop (beef; plus ducks, geese and bobolinks; perhaps chickens), minus the row crops. The agro-ecosystem better retains its nutrients. (One may need some corn and soybeans to “finish” the animals before slaughter.) Raising buffalo on the same land would obviate the need for internal fencing, since buffalo, unlike cattle, space themselves out and rotationally graze a landscape all by themselves. However the cost per mile of the perimeter fence is greater.

Organic agriculture doesn’t use artificial fertilisers or pesticides; so-called sustainable or regenerative agriculture focusses on the whole environment of the farm; and the place of the farm in the larger landscape. There are two long-term sinks for carbon in agro-ecosystems: woody crops and perennial grassland or prairie. Besides storing carbon in their tissues, both trees and grasses transfer it to the soil, where it accumulates. Cattle, sheep, goats and hogs can be raised on grass (the first three are adapted to eat grass, which we cannot). Chickens also can be raised on permanent pasture, in small flocks in bottomless cages that are moved daily; the birds eat grasses, soil insects, and invertebrates. Such birds lead good chicken lives, taste better, and sell for double the price of battery-raised hens. Their eggs contain less chloresterol. In Minnesota, hazelnuts (a perennial woody crop) yield a similar volume of nuts per acre as soybeans. The nuts are a source of proteins and oil and can be managed (pruned, harvested) mechanically. Like pastureland, hazelnuts add to the soil’s carbon stores. They hold soils. Mixtures of native prairie plants can be harvested for their edible or oil rich seeds, though yields are still below those of annual grains. The forage that remains after the seed harvest can be fed to animals. From the perspective of the long term, oak woodland in the hillier, damper parts of the Middle West, interplanted with black walnuts, will slow down soil movement essentially to zero and yield as much net income as any other crop; but unless one seeds the young woodland with morel mushrooms or truffles, or cultivates portobello mushrooms in stacks of thinned oak logs, or grows berries as a understory crop, the harvest won’t come in until one is dead. After the first century or so, the land could be sustainably logged for a net income equivalent to that from row crops ($55 an acre in 1996; so at current prices, one tree per acre every 20 to 40 years). Such long-term conversions would have to be supported by government policy. Support might include loans based on the value of the future crop; annual payments based on the value of the land in the water cycle, or annual payments (subsidized by a tax on fossil fuel consumption) of $20 to $50 a ton for carbon storage. Young, growing temperate forests store carbon at a rate of half a ton to a ton per acre per year and old forests continue to store carbon in their soils. Borders of forest or grassland along streams, as well as permanent prairie or hayland, also accumulate carbon. So do soils under minimum tillage. Soils hold considerably more carbon than the atmosphere. At $50 a ton, carbon storage could be worth a billion dollars a year to American farmers. Evergreens grown for paper pulp, as on the prairies of Minnesota, would provide a faster return than nut trees or natural forest, but since they are clear-cut, and carbon-neutral over a rotation should not be available for carbon payments; however they also restore soils, slow soil movement and turn surface water runoff to underground flow.

Prairie potholes are easily restored wetlands, functionally speaking (one plugs the drain), though, even with assistance, it takes a very long time for their vegetation to return to something like the original condition (if ever: this may or may not matter much to the ducks, amphibians and songbirds of the original landscape, most of which soon recolonize the pond). Pothole wetlands also can be harvested in late June, after the ducks have done nesting, for wild hay. Because the yields in the wet soils are so high (three times normal, up to eight tons per acre), and more reliable year to year than grain, the return from wild hays over the long term is much the same as corn. (One needs a market for the hay.) The wild ducks that once nested in the prairie pothole region are now probably worth more (net) than the grain the land produces, but since the ducks are hunted elsewhere, wild duck farmers don’t receive a return for their crop. In an ideal world, grain growers in the prairie pothole region would combine their farms into larger entities and raise free-range buffalo and elk instead of soybeans and corn, receive a duck subsidy from the sale of shotgun shells, and a carbon subsidy for their carbon-accumulating soils, a payment for keeping excess water and nutrients out of streams, and live comfortably. Forestland and restored prairie pothole habitat is of course land taken out of intensive food production, though both landscapes still produce food (the forestland mushrooms, a high protein food).

* * *

The matter of sustainable agriculture brings up three issues: feeding a world population of 8-10 billion people (that is, yield per acre); profit to the landowner (net income per acre); and the effects of an agricultural practice on its soils and the environment downstream (its biological effects and its sustainablity). The ecological value of undisturbed uplands has been put at $200 to $400 per acre per year. While all such figures are suspect, they provide a basis for discussion. We can say few people are going to pay a farmer that for leaving his land undeveloped, for instance. We can also say farming should try not to extinguish that value. There are few trade-offs in modern human use of the land: only relative losses. Considering agricultural use a trade-off in which the return from agriculture replaces the land’s ecological services is part of the capitalist rationalization in which natural values are methodically extinguished in favor of human economic ones. The loss of ecological services will be felt economically somewhere sometime (usually downstream, several months to several centuries later).

Regenerative agriculture is a step beyond organic agriculture, in that it attempts to regenerate healthy soils and function within a whole working ecosystem. It regards soils as the basis of agriculture and not something that holds plants up. In nature, soils are formed out of their parent material (rock, glacial till) by climate, topgraphy, biotic processes and time. Half the volume of a healthy soil is living material (mostly plant roots). The main biogeochemical transformers in the soil are the soil archaea, bacteria and fungi (the first two groups are what made earth earth); such soil microorganisms constitute 1-4% of soil organic matter but constitute its largest pool of rapidly recyclable nutrients (carbon, nitrogen, phosphorus, sulfur). They provide the nutrients for plant growth, partly by breaking down decomposed plant material (the recalcitrant humus, and other soil organic matter), partly by weathering new nutrients out of the soil’s parent material. Most species of these soil organisms and their functions are unknown. As in a forest, soil fungi expand the root area of crop plants and connect the roots of different plants and species of plants: soil fungi are thought to transfer nutrients from the roots of dying crop plants to those of the germinating new crop, helping the new crop get off to a rapid start. Cultivated plants were once wild plants and regenerative agriculture treats crop plants as though they were wild plants adapted to live in mutualistic relationships with soil micro-organisms and to grow successfully on the pools of organic nutrients in the soil.

Soil nutrients come from incorporating crop residues, legume cover crops, and manures. (Some modern, high-yielding crop cultivars are adapted to very high levels of nitrogen and need artificial fertiliser to perform well.) Some soil amendments, such as ground limestone, are necessary for long-term agriculture. Small applications of lime and phosphorus, together with plant breeding that produced varieties of soybean that would grow in the tropics, opened up Brazil’s Cerrado, the savannah south of the Amazon, to commercial agriculture, with the result that land values there rose 10 times from 1997 to 2004, and that very little of it is left in a wild state. Sugar cane, the world’s largest crop, can be successfully grown without any nitrogen fertiliser if the soil is inoculated with the right species of endophytic bacteria. Adequate levels of soil organic matter are essential for maintaining soil fertility, porosity, and water-holding capacity. Most soils show microbial and fungal activity and have plant roots down 5 to 8 meters, that is well into the soil’s parent material. Alfalfa draws up nitrate from 2 meters down the first year, greater than 5 meters in subsequent years and thus (besides its work in providing a home for nitrogen-fixing bacteria) increases nitrogen levels at the soil surface. (The nitrogen at deeper levels of the soil comes from leaching and will be replaced by leaching from above.) The amount of carbon in a soil (soil carbon largely comes from decomposed plants) is a good measure of its fertility. Studies have shown a 20% increase in potential yield with every 1% rise in soil carbon. A rise of 1% is however an enormous rise: the carbon content of undisturbed soils is about 2%. This falls to 1.5% during the early stages of cultivation and to 0.5% in badly damaged agricultural soils. In temperate regions, incorporating grain straw will increase the carbon and nitrogen in the soil 40-50% percent over 20 years (that is, result in about 0.5% increase in soil carbon). Topsoil can be increased under appropriate agricultures at about an inch a decade (raising productivity about 6%, that is, almost doubling it in a century), while soil production under natural conditions is about an inch every 250-1000 years. Productivity declines caused by soil degradation in conventionally farmed soils have been masked over the last 75 years by planting higher yielding cultivars, increasing rates of fertilisation and irrigation. Yields under soil improving regenerative and organic agricultures are equal to, or about 5% less, than those under conventional agricultures using artificial fertilizers and pesticides. Breeding can produce plants that perform well under high nitrogen regimes, or that are more efficient users of nutrients and water, and thus suitable for regenerative farming schemes.

The nitrogen that flows off farm fields comes largely from artificial nitrogen fertiliser, but also from manure, from nitrogen fixed by legumes and free-living soil bacteria, and from that which falls from the air (fixed by lightning, forest fires, and by industrial and vehicular combustion). The nitrogen fixed by people is now about equal to that fixed naturally. Its effects on the natural world are largely unknown. Nitrogen falling from the air encourages weedy annual grasses in prairies of native perennial grasses in Minnesota (reducing the soil’s storage of carbon) and speeds up turnover of tree litter in urban forests (so one sees few dead leaves on the surface). Streams in unpolluted drainages in South America cycle mostly organic nitrogen. For the last century the natural nitrogen cycle in north temperate ecosystems may have been overwhelmed by inorganic nitrogen from combustion and fertilisation.

At any rate, about half of the nitrogen supplied by fertilisers to crops is used by them. Ten or fifteen percent leaches out with rainfall or irrigation water; some blows away with the soil; up to half (the number depends on the climate, the time and method of application, the type of fertiliser used) is denitrified by soil bacteria and returned to the atmosphere as nitrous oxide. Nitrous oxide is a greenhouse gas that now contributes about 6% of the anthropogenic greenhouse effect; it also depletes the stratospheric ozone layer. A doubling of the amount of nitrous oxide in the atmosphere, which would be a disaster, would lead to a 10% decrease in the ozone shield, which implies a 20% increase in the ultra-violet light reaching the surface of the earth. So conventional agricultural soils lose not only carbon (as carbon dioxide) but also nitrous oxide; thus global agriculture’s atmospheric connections affect both climate and the ozone layer. Agricultural activity is thought to produce 15-25% of current global warming, some from its production of gases like carbon dioxide, nitrous oxide and methane, some from soot and dust, which, falling on Arctic ice and snow, help it melt. The lower number is currently favored. Some agricultural chemicals, like the soil fumigant bromine, now banned, but used by California strawberry farmers under special license, also destroy stratospheric ozone. Soils under no-till agriculture (which reduces or eliminates plowing), store carbon instead of losing it; soils cultivated this way might be able to sequester 8-20% of the carbon emissions in the United States. No-till agriculture also reduces soil erosion and slows water runoff. Conventional no-till agriculture uses chemicals for weed control but since water runoff is less, degradation of the chemicals is thought to be more complete.

The nitrogen that runs off fields enters surface waters, or sinks into groundwater. Some of it is removed by denitrifying bacteria in the damp anoxic lowlands near watercourses, and also in the river waters themselves as they flow downstream. These bacteria return it to the atmosphere as nitrogen gas or nitrous oxide. Continental ground water constitutes 40% of global fresh water, rivers constitute something like 0.005%. Rivers are in contact with groundwaters many times their size and exchange soluble nutrients with these waters. The water in a river spirals downstream, sometimes in the channel sometimes not; the water one sees in a river is a small part of the river-groundwater flow. Riverine invertebrates have been collected from 10 meters down in wells in the floodplain of the Flathead River in Montana up to a mile from the river channel. Such groundwaters act as both a source and a sink for nutrients, and nutrients in the river water can be removed all along its course in riverside wetlands and low banks (as well as by free-floating denitrifying bacteria); more reasons why floodplains and wetlands and aquifer recharge areas all along the courses of rivers should be protected. Since most rivers in industrial countries are overwhelmed with nitrogen, and much of the capacity of riverine wetlands to remove nutrients has been destroyed, most of these nutrients now end up in the sea. The load of nitrogen and phosphorus in British estuaries from farms, industrial combustion, human urine, from fertiliser spread on lawns, from pet faeces, and from land development (losses of nitrogen and phosphorus are speeded up in disturbed soils) is 100 times the level needed to cause eutrophication of those waters.

Nitrogen compounds from farms and industrial combustion also fall out of the air. In large parts of the northeastern United States nitrogen rains out of the atmosphere at the average rate of its application to American farmland. Such excess nitrogen over-fertilises and acidifies forests that are naturally nitrogen-limited, causing a weakening of the mycorrhizal connection between fungi and tree roots, and a reduction in fine root mass (both are adaptations of trees to nutrient-limited lives); the loss of such connectivities makes the trees more vulnerable to droughts and micronutrient deficiencies. The nitrogen also reduces the competitiveness of nitrogen-fixing species of bacteria, including those in lichens. So lichens become less abundant. Calcium in acidified soils leaches out and is replaced by (toxic) aluminum. In native grasslands, the excess nitrogen encourages weedy (often introduced) annual grasses in their competition with the deep-rooted perennial grasses and thus reduces the landscapes’ carbon-storing capacity, which is mediated through the huge root masses of those perennial grasses (ninety percent of the mass of native perennial grasses is underground and one third of their roots die and regrow every year, adding huge amounts of carbon to the soil). Careful hosts in farm country give you coffee that has been made with filtered water, since the ground water (their well water) contains so much nitrogen (as well as various mutagenic and hormone-mimicing pesticides) that it is not safe to drink. Iowa lakes and rivers have some of the highest levels of nitrogen and dissolved phosphorus in the world. A study in England listed the externalities of modern agriculture as the costs to water companies of removing nitrates, pesticides and farm pathogens from drinking water; the costs to companies of workers sick from food poisoning from eating salmonella-contaminated, mass-produced animals; the costs of restoring damaged habitats; the costs of compensating for the air pollution and greenhouse gas production from agriculture; the costs of soil erosion; and the costs of the damage to farmers’ health from farm chemicals. (I find it doubtful the list is complete: habitat fragmentation, introduction of exotic plants, animals, insects and diseases, and persecution of predators come to mind.) Total costs per acre per year were calculated at $120. This cost amounts to a hidden subsidy to modern farming. A hundred dollars an acre applied to all the world’s cropland would make the hidden subsidies to agriculture about the same as the direct subsidies paid to agriculture in the developed world (about a billion dollars a day; $2.20 per cow per day in the European Union).

A cost that study seemed to ignore is that of antibiotic resistance caused by feeding farm animals prophylactic antibiotics. Antibiotic supplements help animals gain weight and deal with the stress of living in over-crowded conditions, with no outlet for their instinctual behaviors (rooting, nest-building, scratching, the formation of social hierarchies, social rubbing, pecking). In the case of cattle raised in feedlots, the antibiotics also help them deal with the consumption of a feed (corn) to which their stomachs are not well adapted. Cattle are adapted to eat grass, which we cannot: one biological reason to eat meat. The acid rumens of corn-fed cattle also seem to encourage the growth of the deadly 0157:H7 strain of E. coli bacteria, which is common in the manure of cattle fed on corn. (E. coli 0157:H7 appeared in 1982 and is thought to have evolved in the guts of feedlot cattle.) The numbers of 0157:H7 E. coli in the manure fall by 1000 times after a week of feeding cattle hay. Genes for antibiotic resistance are common in bacteria and bacteria commonly swap genes among and between species. (Some archaea share genes so often that species distinctions become questionable.) So bacteria exposed to continual low doses of antibiotics quickly develop antibiotic resistance. Bacteria in soils affected by feedlot runoff show considerable resistance to antibiotics. Antibiotic resistant bacteria are also found in manure from dairy cattle, in the air around farms, in wild animals, and in retail meat and poultry. Antibiotic resistance can easily jump to bacteria that affect humans. While it is difficult to trace resistant bacteria in humans to the use of antibiotics in farm animals, antibiotic resistance in bacteria that cause disease in humans is a growing problem. Feeding prophylactic antibiotics to farm animals is foolish, on a level with the feeding of processed cattle carcasses to cattle. Such industrial cannabilism gave us mad cow disease; and probably wasting disease in wild elk and deer (which are also fed food of questionable origin by humans). Such results might have been predicted, since the ritual eating of humans produces a similar neuro-degenerative disease in humans. The eventual economic costs of such practices are likely to dwarf the benefits.

Runoff waters from farms and feedlots contain not only nutrients and antibiotic resistant bacteria but diseases and parasites from the animals. Animals produce about a billion tons of manure per year in the United States (several times the manure of people) and most of this material is not handled well. The disease-causing organisms end up in rivers and eventually in the marine environment. Some of them end up in drinking water and cause outbreaks of disease. Oocysts of Cryptosporidium, Giardia, and Cyclospora (all of which cause human disease) have been found in high levels in bivalves on the East Coast. Cysts of Toxoplasma gondii, a parasite of cats, are washed with cat faeces into the ocean off California. The cysts are concentrated by shellfish. The shellfish are eaten by sea otters, in which the parasite causes severe mortality.

Uptake of nitrogen fertiliser by crops could probably be improved by about 10%, through proper timing of fertilisation, and proper balance of phosphorus and potassium applied with the nitrogen. One problem is that the nitrogen in artificial fertilisers is extremely available—to plants, to leaching by rainfall, to denitrifying bacteria. (It is manufactured to be available.) Anhydrous ammonia becomes available as nitrate at 50° F., before most crop plants can use it. So-called organic nutrients, derived from bacterial decomposition of material in the soil, a biological process that is also temperature dependent, become available somewhat more slowly and at temperatures more in line with the ability of the plants to absorb them. Some fertilisers (especially the ammonias) tend to volatilize rather than leach, so more of them is lost directly to the atmosphere than to surface and ground water. If denitrified to nitrous oxide before volatilization, they contribute to global warming; if volatilized as ammonia, they contribute to acid rain and end up in watercourses by a different route. (Some of the ammonia oxidizes to nitrogen gas and is harmless.) About 50% the nitrogen entering the Gulf of Mexico comes from commercial fertiliser, 15% from manure (mostly from feedlot runoff). Modern, high-yielding, nitrogen-demanding crop varieties need excess nitrogen to grow well. Under good management 35-40% of that nitrogen is going to run off. Some of it will sink directly to ground water, some of the rest will be trapped by grasslands, wetlands and woodlands downstream of the field, but the more nitrogen that runs off, the more land is needed to soak up the excess, and the less effective that land will be. Land used to catch nitrogen is also land taken out of agricultural production (though it may produce hay and therefore meat). Fertiliser use comes up against limits. By ruining riverine and ocean fisheries, intensive commercial agriculture may reduce the total amount of food a watershed produces. One can’t maximize grain yields and have a healthy aquatic environment; but one can have lesser amounts of grain, together with grass and hay (that produce milk, eggs and meat), some timberland (that produces logs and game animals), and riverine and marine fish. Such changes probably imply a smaller human population.

Organic agriculture takes land out of grain production through its rotations into legume sods. Modern crop plants have a much greater proportion of the plant in seeds, rather than in leaves and stalks, compared with traditional varieties. Thus recycling their residues replaces less of the nitrogen and phosphorus lost in the crop than formerly. (One very competent writer claims that a complete recycling of all the organic matter from currently harvested land and from all confined animals would not replace the nutrients lost in crops under modern high-yield farming.) Spreading manure produced by humans, capturing the nitrogen in human urine (freezing sewage effluent is one of the more off-the-wall suggestions: since urine freezes at a lower temperature, the excess water can be removed as ice, methane from the sewage would power the process), and composting food waste would make most of that up. Adding rock powders and rotating land into sods also help. Legumes (alfalfa, vetch, clover) fix considerable amounts of nitrogen (300 to 500 pounds per acre), but only 5-10% of this is available for the following crop. The rest adds to the soil pool and slowly becomes available over time. On a regenerative farm, rotating land into legume sods, incorporating crop residues, rotating grain crops, and adding composted manure provide sufficient nutrients for a crop slightly less or equal to that of one’s high-tech neighbors in a good growing year, and considerably more in a drought, when regeneratively farmed soils perform better (probably because of their higher fungal populations, and better mycorrhizal connections with the roots of crop plants). Good regenerative practices raise levels of soil carbon; expand the labile pool of soil nutrients (that more available for growing plants, including the fresh plant material subject to rapid decomposition); enhance nitrogen fixation by soil bacteria; reduce soil erosion; and improve soil tilth, porosity, and water-holding capacity. (All this is work done by soil micro-organisms, fungi, and invertebrates.) Soil tilth—its crumb structure, which strongly affects its water holding capacity—is created by gluey polysaccarides secreted by mycorrhizal fungi. Nitrogen fertilisation makes crop plants withhold exudates from the fungi, negatively affecting soil tilth. The ground freezes less deeply under fields in better tilth, letting the farmer plant earlier in the spring. The water that flows off the land is cleaner. Crop breeding has raised corn yields by 31% since 1995, 76% since 1975. These crops were bred to be fertilised with artificial nitrogen. But crops could also be bred to take advantage of symbiotic soil bacteria and fungi, that is for high yields in an organic agriculture; and to withstand drought.

Use of manures and sludges (human manure plus other things that go down the drain) have their own problems. Both animal manures and sewage sludges contain heavy metals (including cadmium, arsenic, chromium, lead, mercury, nickel and vanadium). Those in animal manure originate partly in synthetic fertilisers (especially phosphate fertiliser). Some of these could be eliminated. Other metals come from plants. Some plants have an affinity for metals, which can make them useful in cleaning up contaminated ground. Alfalfa plants for instance, will sequester gold, while aspen trees will take up mercury and volatilize it from their leaves, thus distributing it more widely in the atmosphere (how much this helps is another matter). Food plants with the metals in them (perhaps metals falling out of the sky from coal burning power plants) are eaten by animals and the metals come out in the manure. Sewage sludges contain metals from the plants and animals we eat, as well as from roadways, rooftops, factories, and labs (such as metal-plating or film-developing facilities). So, if they are to be used as a soil amendment, sewage sludges must be cleaned up; and matched to the soils they are spread on; and the rate of application of both sludges and manures limited, so as to avoid build-up of heavy metals in the soil. Cleaning up urban sludges is partly a matter of controlling what goes down the drain (capturing metals at the source); and partly a matter of better processing. For instance, holding the material in tanks for two weeks rather than for four days, and shifting between tanks with aerobic and anaerobic fermentation allows bacteria to remove more nitrogen from the effluent and seems to break down chemicals like estrogen; estrogen is probably not a problem for material spread on land but is a problem with sewage water released to waterways, where the hormone feminizes fish. (Some of these fish end up on the supermarket shelf; and the estrogen interferes with fish reproduction and development. Estrogen added to an Ontario lake in concentrations similar to those found below sewage outlets in rivers caused a 2000% drop in the local minnow population in two years.). Holding material for longer periods means more tanks, more capital investment, and better management. There is no question that human and animal manures can be used safely as a fertiliser: some Chinese landscapes have been farmed for 7000 years using human manure.

Some manures are made toxic from drugs and pesticides fed the animals. The chicken feed used to raise chickens (one billion a year) on the Eastern Shore of Maryland contains arsenic and anti-biotics. The arsenic is a biocide used to control parasites in the chickens, the antibiotics keep the chickens healthy and make them grow faster (it improves their use of feed). Excrement from the chickens is used to grow corn and soybeans for chicken feed. So arsenic now contaminates surface and ground waters along the Eastern Shore and the soils contain antibiotic resistant bacteria. Eliminating arsenic and antibiotics from feed may mean changing the way chickens are raised.

Manures and sludges are heavy and expensive to move, but using them on farms is undoubtedly the best way to use them. Thus large animal facilities worsen the disposal problem. A dairy farm with 1000 cows, whose animals produce faeces equivalent to a village of 10,000 people, needs 2000 acres to spread the material; somewhat less if the farmer puts it through a methane digester or composts it. (Both these processes will reduce its pathogens virtually to zero.) Most confined animal operations store their manures in open lagoons, where organisms that cause human disease are not appreciably reduced and where anaerobic decomposition produces considerably more carbon dioxide and methane than aerobic composting. During heavy rains such lagoons often fail, and send their nutrients, in toxic amounts, into rivers. Most confined animal operations have far too little acreage to spread their manure. Their investment is in their facilities. Phosphorus especially is a problem. Phosphorus is often the limiting nutrient in rivers and lakes and adding it causes algal blooms and eutrophication. The ration for the animals in a confinement barn or a feedlot comes from grain raised a hundred or a thousand miles away, and the animals produce too much manure for the surrounding fields to absorb. Only 15% of farms in the Corn Belt currently spread manure.

In general, while organic, soil regenerating farming produces yields slightly lower than modern chemical agriculture, it yields higher returns to the farmer. This is partly because the price the farmer receives is higher, partly because purchased supplies are less. But yields per acre are not only dependent on fertilization; they also depend on general farm practices. Planting spring wheat, corn or soybeans two weeks earlier in the short growing season of Manitoba boosts yields from 10-40%. While good management of crop residues helps prepare the land for early planting, reliable early planting usually means drainage (and then appropriate handling of the water that comes out of the drains). Careful management of crop residues enhances nitrogen fixation (providing nitrogen for the next crop), reduces compaction and erosion, and improves the soil’s water holding capacity—so the soil freezes less deeply. In California’s Central Valley a study compared tomato yields with organic and conventional methods. Without the use of conventional pesticides and fungicides, yields were predicted to fall 36%. But chemicals turned out to be less important than inherent soil fertility, water availability, the cultivars grown, and management practices (such as transplanting time, which strongly influenced the degree of insect damage). The density of herbivorous insects was only 5% greater on organic farms, but the herbivores were considerably more diverse (they belonged to more species). The predators and parasitoids of the herbivores were much more diverse and much more abundant on the organic farms.

In some cases, pesticides reduce total yields of food. Farmers in Bangladesh who stopped putting pesticide on their rice so they could grow fish in their paddies, found their yields of rice rose 25%; and they also were able to harvest the fish. One reason for the increased rice yields may have been a speedier cycling of organic nutrients by the fish; part of a generally more efficient food web in the paddy that had been suppressed by the pesticides. Aquaculture in the paddies, by withdrawing nutrients from the water, means the water flowing off the paddies is cleaner. In China, rice yields have been increased by draining the fields several times during the growing season. This saves water and also reduces methane production in the rice fields by 40%. (Of course, such draining eliminates the possibility of raising fish. Methane is another greenhouse gas.) Some rice farmers in the Sacramento Delta of California now flood their rice fields after harvest instead of burning the rice straw; the straw decomposes in the water, improving the organic content of the soils, and migratory waterfowl use the fields as a substitute for natural wetlands. The waterfowl recycle in-field nutrients and add more nutrients from their feeding in upland habitats. (Such cycling can be substantial. It is thought that nesting seabirds contributed two million tons of guano annually to New Zealand coastal forests, a thousand pounds an acre, before Pacific rats brought by the Maori destroyed their nesting colonies.) Some of the California rice fields are leased to hunters, producing additional income, but overall such management is a good deal for the ducks and geese, whose winter numbers in California have fallen from 100 million to perhaps 6 million over the last century, as 90% of California’s wetlands were drained. (Large parts of the Central Valley were marshy in winter and spring; winter wheat can be grown there without irrigation.) Rice farmers in Louisiana grow crawfish, and also provide habitat for native waterfowl, in their paddies during the off season; the crawfish eat the vegetation that grows in the paddies.

In Chinese cities urban gardens provide over 85% of the fresh vegetables used in the city, plus some meats, grains, and tree crops. On the outskirts of Italian cities one sees gardens on the slopes of drainage canals. Such hand agricultures are always more productive than extensive cultivation. Farmers in the United States now need 30,000-45,000 square feet to feed a person on a diet high in beef, considerably less if more of the animal products are from dairy cattle, poultry, or pigs (which are more efficient converters of grain to meat or milk), 10,000 square feet for a vegetarian diet. Such numbers vary; these figures are for single-cropped, chemically farmed cropland, and include the large amount of waste currently in the American food supply (about half the food produced is wasted at the retail level). Such farming is energy intensive; on average in the United States, food contains 0.1 calories of the fossil fuel energy used to grow it. On the other hand, a bio-intensive gardener needs 2000-4000 square feet (20-40% of the land of the commercial farmer), seeds, and hand tools, to feed a person on a vegetarian diet, with a small amount of chicken or pork. This includes land for compost crops. In this case the food contains several times more calories than the work that went into it. (Among the most productive hand agricultures in the world are the dike-pond units in the Pearl River delta of China’s Guandong Province. Irrigated agriculture and aquaculture are integrated. The nutrient-rich runoff from the fields fills fish ponds, which are also given crop wastes. Periodically the ponds are drained to grow crops. Crops include cows, pigs, vegetables, mulberries, sugarcane, ducks and fish. This organic agriculture yields 30 to 50 tonnes of fish and crops per hectare, or enough to feed 50 or more people, while intensive rice farming in the same region feeds 11, as would American commodity agriculture. Several different carp species with different feeding niches are raised: surface feeders, middle layer feeders, fish that eat plants and organic wastes, bottom feeders.) Figures like this have led a geographer to quip that to improve agricultural yields we should suburbanize the landscape, and let people grow their food in gardens. In cities with appropriate climates and sufficient capital, rooftop gardens under greenhouses made of photo-voltaic panels (some are now transparent), could produce electricity and grow vegetables, fish, chickens, and rabbits, a variation on the rural schemes of the New Alchemists.

Feeding the human world, even a world of 10 billion people (though that number now seems less likely), is less a matter of improving the yields of farmland (though in some cases that would help) and more a matter of focus on the problem of population, diet, agriculture and landscape as one whole. In 1948, 50 million pounds of pesticide were used in the United States and 7% of crops were lost to pests, preharvest. In the 1990s, 20 times that weight of pesticide was used and 13% of preharvest crops were lost. Some of this is due to the fact that plants have been bred for yield rather than for insect residence (sprays take care of the insects), some from the planting of the same cultivars on hundreds of thousands of contiguous acres. With their short generation times, insects and plant pathogens adapt to changing conditions much faster than plants, so such plantings make the plants very vulnerable to a single adaptation on the part of the pests. Some of the increase may have been due to the growing size of fields, the increasing use of the whole habitat for crops, and spraying for weed control, all of which made the environment less favorable for insect predators of pests. Most agricultural chemicals are bad for people and animals, and make fields and orchards dangerous places for wildlife and farmers. They kill insects that birds and mammals need for food. In sufficient doses pesticides directly kill birds and mammals (60-70 million birds a year in the United States), as well as amphibians and other vertebrates and invertebrates. When predatory animals eat the dead insects, the dose is concentrated in them, so orchards are not healthy places for predatory insects, or for small falcons that eat large insects, or for red-tailed hawks that eat mice sprayed with organophosphate pesticides. In very low doses many agricultural chemicals weaken the immune systems of amphibians and mammals, cause or promote cancer, feminize developing embryos, cause mutations. Atrazine, a weed killer used extensively on corn, is an immuno-suppressant. It is detectable in Middle Western brooks and in spring rains over the Northeastern United States. Most of what has been applied in the Middle West is still in the Great Lakes. Besides its immuno-suppressive activity, in very low doses (3% of the Environmental Protection Agency’s standard for drinking water) atrazine makes male northern leopard frogs hermaphroditic: so populations of northern leopard frogs in agricultural environments are deformed (deformed by parasites their immune systems can no longer handle) animals of uncertain sex; and their populations have plummeted. Atrazine interferes with the formation of shells in mussels. Many chlorinated hydrocarbons are toxic at two levels, one high, and one very low, where their concentration approximates that of hormones in the blood. (The higher concentrations are apparently ignored by the body’s signaling systems.)

While spraying is sometimes necessary, there are also other ways of controlling pests. Codling moth is a major pest of pears, apples and walnuts in California. Pear orchards suffer 5% damage or less when a bat colony is less than a mile away; and 60% damage if the colony is over 2 miles away. Large bat houses were once built in the American South as a method of mosquito control. Building bat houses in orchards (or installing boards for bats to roost under along the sides of barns) would be much cheaper than spraying. And one could spread the guano. An organic farmer in Oregon controlled the insects in his cherry orchard by hanging nesting boxes for swallows on wires among the trees; the birds ate the insects, and when he cleaned out the nest boxes in the fall, he used the nesting material to fertilize the orchard. Swallows and bats are limited more by roosting and nesting habitat than by food supply, so populations can be increased (at least locally) by providing nesting and roosting sites. Similarly, shrike populations can be increased by providing perching posts. Shrikes eat small birds, large insects and mice. Citrus growers in China place colonies of predatory ants on their trees, and dig moats about the trunk to keep them there; such trees have 60% less insect damage than sprayed trees: the ants do a better job than the spray. (Ant colonies inhabiting acacia trees in Africa swarm to protect them from browsing by elephants and giraffes—so some trees are left after the elephants have moved through.) Wild strawberries and roses in orchards in the Pacific Northwest provide winter habitat for wasp parasites of leafrolling insects, keeping their populations high enough so in spring the wasps are there to protect the trees. Owl nesting boxes and hawk perches are used to attract predatory birds to control mice in organic vinyards in California. Some organic orchardists graze pigs among the trees. The pigs keep the grass and weeds down, and eat the fruit from the June drop, killing the larvae of the plum curculio, a major pest of apples, thus preventing them from migrating into the soil and developing into adults. Modifying the structure of the agricultural landscape provides habitat for predatory insects and rotating crops breaks population cycles of pests. Here we are looking at things like smaller fields (pollinators are more effective within 100 yards of their nests, predatory insects are more abundant at field edges); replanted hedgerows; in-field wetlands; strips of crops of different heights and leaf characteristics (providing a more varied habitat for predatory and prey insects); several varieties of the same crop in a field (not one cultivar); regular crop rotations; very little spraying: all this against a background of a semi-natural environment (a later successional ecosystem) that occupies 15-40% of the agricultural area (including some prime land) and provides natural habitat for insects, birds, mammals, amphibians and invertebrates (many useful to the farmer) and also serves to regulate the quality and flow of run-off water. Agriculture is always full of disasters, climatic and otherwise; different crops provide insurance. By some estimates there are 5 billion acres of degraded and abandoned agricultural land worldwide (current cropland is about 3.75 billion acres, with 7.5 billion acres in pasture and rangeland). With sufficient investment (say, $100-$200 an acre), this land could be rehabilitated. Successful rehabilitation under regenerative agriculture would produce food, oils, fiber and habitat; the soils might absorb half or more of the 6-7 billion tons of carbon the human world emits annually. (And continue to absorb the carbon for some centuries: soil carbon levels seem to rise linearly with the amount of recycled crop residue added to them, and continue to absorb carbon indefinitely.)

We cannot feed the present world population on a modern western diet. But while probably 15% (about one billion people) of the global population is currently underfed, we produce more than enough food to feed them. Farming is a business. Animals eat 35-45% of the grain grown worldwide, thought the grain would yield more calories and protein if people ate it directly (it’s more profitable to feed it to animals). Quite a high percentage is wasted before it reaches the consumer. (About 10% of grain is wasted in developed countries, up to 20% in underdeveloped countries. Spilled grain in the United States supports Canada geese about Chesapeake Bay and other waterfowl in the rice fields of California and Louisiana.) So grain provides 20-30% of energy intake in affluent countries, compared with 60-80% in poorer economies (and bread consumption in France went from 600 grams per person per day in 1880 to 170 grams per person per day in 1980, the difference being made up by grain processed through animals—eggs, cheese, meat). Global milk production is now about 80 liters per person per year. Traditional milking societies such as those of northern Europe consumed the equivalent of about 100 liters per person per year. Much of the world population is lactose-intolerant and can’t digest milk, but lactose-intolerant people can digest yogurt and cheese. Global meat production is about 75 pounds per person per year. This is also enough: a reasonable annual meat consumption, based on an evolutionary understanding of people’s needs, is about 40 pounds, or what the Iroquois ate. Western societies now consume the equivalent of 300 liters of milk and 150-240 pounds of meat per capita, probably several times that needed for people to reach their full genetic potential for growth. (In foraging societies, one of our better benchmarks for diet, plant and animal protein provided about one third of food energy. People got more vitamins and minerals, including calcium, from wild fruits, nuts, roots, tubers and legumes than currently recommended and much more fiber—beet and cane sugar now provides 20% of the energy in a modern diet. Foragers also took in much less sodium and much more potassium than sodium.) In the West, about 40% of food is wasted after being purchased. Some waste of course is unavoidable. The current problem is less food supply than food distribution: the poor (in global terms) can’t afford food; and their countries either can’t afford to subsidize their agricultures or don’t choose to.

Feeding more people is also a matter of diet. It is much more efficient for people to eat grains, than for them to feed the grain to animals, and eat the meat. Grains are virtually equal in energy density to meat (they have 5 times the energy density of tubers like yams or potatoes), and are not that low in protein: rice is 7% protein, durum wheat 15%. Mixing them with legumes makes a diet balanced in protein; and eating fruits and vegetables provides vitamins and minerals. However there are good arguments for including some meat in a diet. People have a much shorter gut than our close relatives among the primates. Some writers have proposed that (energetically speaking) reducing the size of the human gut made our large brains possible. Cooking, which doubles the carbohydrates available in plant foods, would have been the adaptation that made a reduction in the size of the gut possible (smashing foods before eating also helps make them more digestible). At any rate our short guts adapted us to energy-rich foods: nuts, seeds, meat. If we ate leaves like gorillas or chimpanzees, we would have to spend much of the day eating and digesting, the fate of the once-carnivorous pandas that became vegetarian and now spend most of their days eating the tips and leaves of bamboo. Meat is not necessary in a human diet. But herbivorous animals, thanks to the microbial symbionts in their guts, can eat foods that we cannot (grasses, twigs, leaves: black bears eat deer droppings and thus acquire their bacterial load, which lets them digest grass); and soil-regenerating grain-growing agriculture is not possible without rotations through legume or legume-grass mixtures. The system works better both economically and biologically if something eats those hays, converting them to meat and manure, which can be spread on the grainland in the years after the rotation from hay. (Of course hays can also be used to make paper or biofuels.) Some areas, like the North American plains, are adapted to grass and grazers. Grasslands constitute 20% of the modern global landscape and, helped by the grazers, store carbon in their soils. Using (some of) them to produce meat is far better than using them to produce grain. Many writers claim that an organic, soil improving agriculture, located within a living natural landscape, cannot feed the existing world. The necessity to rotate grains into sod crops (thus reducing total grain production) makes this so; but such calculations ignore all the degraded land that could be brought into production. At any rate, at this point, human hunger is essentially a choice (largely a choice made by the rich nations): like the Irish potato famine, partly a matter of income, partly a matter of eating habits, partly made for us by the invisible hand of the market.

No terrestrial agriculture will be able to feed a population that grows forever. Earlier peoples, without our notions of progress, limited their populations, or starved. One could argue that a population reaches its limits when it no longer fits within a working natural landscape. If so, we reached our limits several thousand years ago, when there were perhaps 100 million people worldwide and agricultural gases began to moderate global climate. (And Deep Ecologists, who say 500 million people are enough, would be right. However that first mild warming was hugely advantageous to us, and more or less a toss-up to the rest of the biological world.) Such matters are fluid. The falling birthrates of the West (the U.S. population would be falling too without immigration) are regarded as a disaster, but if we could figure out how to get through the demographic squeeze (an economic problem), they would result in a much healthier landscape. (If every fertile woman restricted herself to one child, population would fall to 1.6 billion by 2100, a level that would allow a modern life life for everyone. At two children per woman, population would still fall, but more slowly.) It was long thought that development (a certain level of affluence) was necessary to bring population growth under control. But what seems to matter more is improving the survival of children and improving the status of women. If women are educated, can support themselves, and see that their children are going to survive, they—if they can—will limit their fertility. (Not always: among the more fundamentalist Christians, Muslims and Jews, and I am sure amongst some other religions, population growth is regarded as a good in itself.) Populations then stabilize at much lower levels of income. All this of course is in direct conflict with the capitalist doctrine of continuous growth in population and income, one constantly driving the other.

* * *

A sustainable life for a large human population in a working biological environment means dealing appropriately with nutrient flows, whether this is “economic” or not. In Dutch agricultural districts, where large numbers of animals have been raised for decades on imported grain (the Dutch invented raising animals in confinement), the animals’ manure is now dried and turned into a bagged fertiliser that is exported from the region. This is to keep the nitrogen levels in groundwaters, already high, from rising further. In broad economic terms, it is certainly cheaper to subsidize the costs of dealing with the waste from the animals than to deal with groundwater too toxic for people or animals to drink, so all water has to be filtered, or with the eutrophic bays, rivers and lakes into which the water drains. Not all nitrogen fertilisation is harmful. There are degraded soils such as those of the dry Ethiopian highlands where some artificial nitrogen fertilization is biologically helpful. Fertilisation doubles or triples yields of wheat and tef, so more land can go into cover crops and trees; thus improving the habitat and reducing soil erosion. In the dry wheatfields of the Dakotas and Saskatchawan little nitrogen leaching from rain occurs, so nitrogen fertilisation will be less harmful to watercourses, but nitrogen will be lost to the atmosphere as nitrous oxide, and along with soil and phosphorus, to the wind. One has to look at the effects on a place. The point is to farm appropriately, both globally, and in a given environment.

Not including ecological costs in economic costs subsidizes the destruction of the natural environment. Environmental protection then seems expensive; but reducing the nitrogen load in rivers by upgrading sewage treatment plants that empty into them will also raise property values along those rivers. Upgrading sewage treatment plants and subsidizing the spreading of New York City sewage sludge on Texas cotton fields (where its heavy metal content doesn’t matter), or on strip-mined lands in Pennsylvania (where sludges help, and turn wastelands into carbon-storing ecosystems) make economic sense in the long run. (For one thing, abandoned lands become productive.) We must subsidize the appropriate uses of sludges or include such costs in sewage rates. How to pay for such matters is a difficult question, since the benefits of cleaning up, say, Long Island Sound, accrue to more people than those in the sound’s sewage districts. And not just sewage influences the health of the sound; all the runoff in its watershed does, including that from septic systems, roads, cattle, factories, dogs and cats. We probably should subsidize the restoration of natural wetlands (or the construction of artificial ones) to filter and slow the water that runs off the 50 million acres in the Corn Belt that have been artificially drained. Drained land constitutes about 25% of the landscape. Letting 15-25% of the landscape revert to natural lands, by planting it to oakwoods or prairie, by blocking the drains in prairie sloughs, or opening levees to let rivers occupy old wetlands, restore much ecological process to the landscape. New, inexpensive drainage structures that allow the farmer to control the level of water flowing from individual fields may reduce the size of the filtering wetlands needed and also, by allowing the farmer to control subsoil moisture better, raise yields. Lands in the Conservation Reserve peaked at 36.8 million acres in 2007, 8% of the agricultural landscape, at a cost of $1.8 billion (about $20 per acre). Putting 25% of the agricultural landscape in permanent conservation lands would cost perhaps $6 billion a year. Payments for conservation lands would replace price supports for grain, at less cost. (Payments would have to rise if the price of grain rises.) In the coming world natural lands that store carbon will be eligible for other payments. Replanting marsh grasses and cattails on farmed peatlands in San Francisco Bay stores 24 tons of carbon per year. Continuing to farm these lands releases about 7 tons per year (oxidation of the peat has lowered peat islands 15 to 20 feet below water level). At $30-$50 dollars per ton, farmers could net around $1000 per acre. Returns in the Corn Belt would be much less ($30-$50 per acre), those in the sugar cane fields of the Florida Everglades comparable to those of the peatlands of the west coast..

The runoff of water and nutrients from farmland is at least partly a matter of economic attitude. In the Pearl River Delta of China, such runoff water would be regarded as useful, along with the ponds, marshes and woodlands used to filter it. In the Middle West, in a better managed landscape, the excess nutrients would grow timber, marsh plants and fish. But much water pollution is pointless. Lawns occupy 40 million acres in the United States. The fertiliser and pesticide used on lawns and gardens amounts to about half that used in the country, and virtually none of it is economically necessary, except to the fertiliser companies: compost, produced from yard waste at the local landfill and distributed free, will produce a healthier lawn and garden. Use here is a matter of economics (cheaper to spread), esthetics, and attitude (moral notions), which come together in property values: a green lawn is considered the “right” setting for a house, and fertiliser and pesticides as “clean” materials. Land uses based on moral attitudes are harder to deal with. (“Picking up” the environment is another one: natural environments often appear messy and chaotic; but the tangle of branches and trees lying on the floor of a natural forest—the coarse woody debris of the foresters—retains moisture during dry periods, provides sites for nitrogen fixing bacteria, and represents a pool of nutrients for the forest. Its removal may help explain forest decline in central and eastern Europe. The fallen logs and branches are an important part of small mammal habitat, providing dens, cover and food plants like mushrooms. They help protect tree and shrub seedlings from browsing deer. Healthy small mammal populations help maintain the ecological processes of which they are a part; forest voles disperse the spores of mushrooms, whose mycelia help maintain the trees; voles are also food for predators and—predators themselves—help control defoliating insects.) Even without lawn fertilisation, water that flows from suburban settlement will be polluted (by animal faeces, metals, motor oil). Great rivers of underground water flow below most agricultural and urban areas, all susceptible to pollution by leaching from above. The water in the Ogallalla Aquifer under the Great Plains moves at a rate of a few feet a year from the Rockies toward the Mississippi Valley. One could argue that that movement, which brings it across private property and state boundary lines, makes it public property; its waters also feed surface waters, which, if connected to navigable waterways, come under federal regulation. (Several states have declared groundwaters a public trust and thus subject to regulation.) Underground waters are far too essential for human life (and for the biological world) to be polluted; or to be depleted beyond their natural rates of recharge. Once polluted, they are astronomically expensive, at least with current technologies, to clean up; once depleted they are gone: where it was least abundant, the economically extractable part of the Ogallala water is already gone.

When the snow melts in the spring, one sees in the bare fields the shadows of its former hydrology: damp spots that were once depressions in the forest or prairie, the dark lines of temporary streams. Such surface hydrology, with its plants and animals and their water cleansing effects, is what the famer eliminates. To a certain extent, the release of agricultural nutrients to streams and the atmosphere, can be controlled by agricultural practice: what fertilisers are used and how they are applied; how much manure is applied and how it has been handled; what crops are grown; how fields and hedgerows are laid out; how crops are rotated. Physical constructions also help. They amount to a sort of reconstruction of the hydrological landscape: vegetated corridors along streams, constructed wetlands for drainage water, grassed drainage channels from fields, permanently vegetated aquifer recharge areas. Some of this can be profitable: one farmer in the prairie pothole region who restored his wetlands under an agreement with a conservation organization (and then kept them restored because he liked the ducks), found that cutting the slough hay when the ducks were done breeding produced about as much net income as growing corn, partly because he didn’t have to seed the hay and partly because the hay yield was more dependable than the corn crop. (He fed the hay to his young cattle.) Dabbling ducks like mallards, pintails and teal prefer shallow seasonal wetlands, with their abundant aquatic life, in early spring. These can be temporary pools in a thawing field. They need somewhat deeper ponds later in the season for nesting. Such ponds provide cover for the ducklings and an abundance of the small aquatic invertebrates the ducklings need to grow. (In modern agriculture, pesticide drift can reduce these invertebrates by up to 90%, reducing the value of the habitat to zero.) If a farm landscape is varied and productive enough, with small grains and hayfields where a nesting can be completed, game bird populations allow a farmer to lease hunting rights. In the Midwest and Rocky Mountain states most of these birds will be introduced ones (ring-necked pheasants, chukar partridge, hungarian partridge): the native prairie chicken of the prairies tolerates agricultural landscapes poorly, the sage grouse (once a keystone species of Great Basin sagelands) hardly at all. But the new birds become part of the new landscape; goshawks, peregrines and foxes, as well as hunters, dine instead on ring-necked pheasant.

In the end, controlling nitrogen in streams and estuaries means controlling its use. It is virtually impossible to control all the nutrient run-off from commercial cropland: the agro-ecosystem doesn’t work that well and the surrounding nutrient sinks can only trap so much. Levels of nitrogen in Iowa’s Des Moines River were already high under the traditional crop rotations before the Second World War. Plowing, and cultivating for weeds (up to 7 times a season), were causing considerable soil erosion. To reduce nitrogen in streams one has to use less of it. One way to control nitrogen use is by taxing it. Nitrogen currently represents about 10% of a farmer’s cost of producing a bushel of corn. Regenerative farmers that rely on tight nutrient cycling for fertility have an incentive to keep nitrogen out of streams and in the field (through the use of cover crops and crop rotations, through no-till farming); raising the price of nitrogen gives the conventional farmer the same incentive. (To do so effectively, they may have to adopt some regenerative practices.) Reducing the dead zone in the Gulf of Mexico requires a reduction in nitrogen runoff of 40%. So application rates must be reduced by something like 40%. Doubling the cost of nitrogen would go a long way toward doing this: one could raise taxes on nitrogen by 5% a year for 20 years, with the taxes paid by the manufacturing companies (effectively, the price of fertiliser would rise). The monies collected would be used to restore riverine wetlands or construct nitrogen traps on farmland. (Some riverside lands are only farmable because of groundwater pumping during wet periods, usually spring through early summer; these are obvious candidates for restoration.)

A market approach offers a payment. Currently western governments spend over $100 an acre on agricultural subsidies; net profit for the American farmer is about $55 an acre; the ecological value of farmed uplands is $200-$400 an acre. Suppose farmers received a payment of $50 an acre for the ecological value—the public goods—of their lands. Most of this (say 75%) would go for the water that runs off the land: its total amount, the rate at which it runs off, the amount of soil it carries with it, its chemistry (including its load of nutrients and pesticides). A smaller part of the payment would credit the land as habitat: the size of its fields, the number of its hedgrows, the variety of crops grown, connection of its natural lands with neighboring natural lands. To receive the full per acre payment for water, the water runoff and chemistry would have to approach the original condition: rates of runoff, temperature, and chemistry would have to be within (say) 20-30% of pre-settlement levels. The farmer would decide how to do this. A scheme like this could be administered by existing agricultural bureaucracies. Along with Conservation Reserve payments, It would more or less replace price supports. The value of the land as habitat (for pollinators, for insects and animals that help the farmer, for native wildlife) would also be rated for payment.

If one is attempting to lower nitrogen use, one should stop subsidizing its use. Current price supports support nitrogen use by rewarding production alone. The government essentially guarantees it will buy all the grain that is produced, that is, government payments will make up any difference between the world price for grain and what the government calculates it costs the farmer to produce it. Such costs are based on those of modern chemical agriculture. The purpose of price supports is to provide cheap food. Price supports on grain hurt farmers that grow grain to feed their animals by making grain artifically cheap. (Since the costs of raising grain are greater than its market price, which is driven down by oversupply, it’s cheaper to buy grain than to raise it). Thus price supports help feedlots, which are major polluters, by making it cheaper to buy grain (grown on farms without animals, fertilized with artificial nitrogen) than to raise it. An unintended consequence of the boom in ethanol made from corn—a bad but politically irresistible idea—is that the price of corn is rising sufficiently to make the cost of beef raised on grass look better.

Stream health can be assessed in a variety of ways. The abundance of top predators is one way; dippers are birds that occupy linear territories along streams and feed almost exclusively on the invertebrates of the river bottom (the river benthos). Their breeding abundance shows a significant correlation with the abundance of benthic animals like mayfly nymphs and caddis fly larvae, thus their abundance is a way of measuring pollutants like nitrogen and acid rain (which reduce the abundance of these organisms). The breeding abundance of dippers has been used to assess the health of Scottish rivers. Similarly, abundant dugongs and manatees indicate healthy tropical coasts; abundant grassland birds indicate a healthy soil biota in the farmer’s pasture (as well as an assessment of his management practices); abundant seabirds and whales indicate healthy oceans; and sea turtles, large groupers and sharks healthy coral reefs. Of course, aquatic invertebrates, nutrients and pesticides can also be measured directly.

Easily erodable farmland (sometimes streamside or riverside land) is now voluntarily taken out of production through the Conservation Reserve Program. It is cheaper to buy such land outright (usually 10 years or less of payments amount to the price of the land); but outright purchase raises the questions of trespassing (people able to walk into a farmer’s property through public lands) and of local property taxes (which farmers pay on such lands: essentially Conservation Reserve payments shift the cost of property taxes to the federal government). Outright purchase also reduces the farmer’s income. Another method is to purchase permanent conservation easements; in many ways this is equivalent to buying the land, but avoids the problem of trespassing and the matter of property taxes (which however may be reduced on such lands). While underfunded (conservation programs amounted to 8% of federal farm outlays in 2000), probably because it helps small rather than corporate farmers, the Conservation Reserve Program has been a great success, both in reducing soil erosion and in restoring habitat for native animals. (Soil erosion was down almost 40% from 1978 to 1997, although 29% of fields were still excessively eroding. While some of this was due to land in the Conservation Reserve Program, farmers since the early 1980s have also been required to adopt some soil conservation measures to remain eligible for commodity payments.) When the prairies were first settled, the populations of birds like prairie chickens increased (the spilled grain was a resource, as may have been the insects and weed seeds in the fields); as agriculture occupied more and more of the landscape, the birds declined. In hilly parts of the Palouse wheat growing area of the state of Washington, whose deep, light soils are easily eroded, lands in Conservation Reserve (up to 25% of lands in some counties) have slowed erosion and led to a major comeback of the sharp-tailed grouse, a game bird whose subspecies in the Palouse was nearing extinction.

The National Science Foundation has estimated that restoring half the wetlands in the continental United States would require 3% of agricultural and urban land. (About 7% of the lower forty-eight states lies within 100 year floodplains.) Wetlands, existing or restorable, that perform useful ecological work (reducing floods, removing nutrients), and aquifer recharge areas (which filter water) are already known or could soon be identified by government bureacracies already in place (the Corps of Engineers, the Soil Conservation Service, the Agricultural Extension Service). A long-term appropriation (say, for 100 years) of $10 billion a year would give each county in the United States approximately $3 million a year to spend on land or conservation easements. Such monies add up over time. They trade private ownership of land and resources for public amenities. (While one can argue that ecological processes cannot be owned and therefore should, like rivers and lakes, be under the control of the society as a whole, they can be interferred with—by activities such as agriculture and settlement in floodplains—and at this point the only way to get a market society to improve the situation is to make improving it profitable.) Counties in the Southwest might buy up the rights to groundwater reserves whose depletion (by agriculture or development) is leading to the loss of surface flow in their rivers. Such lands can be put together into wildlife reserves and corridors. Recently the Corps of Engineers bought the development rights to 8500 acres of floodplain wetlands along the Charles River outside Boston, Massachusetts; the development rights cost 10% of what a dam would have. Acquiring floodplain wetlands is a permanent solution to the problem and raises the value of developed adjoining lands. A flood in 1986 in Napa, California, caused $100 million in property damage; 5,000 people were evacuated and 3 died. Downtown Napa had suffered from floods for decades. Rather than strengthen its flood defenses, the city decided to purchase 650 acres of leveed-off wetlands and move some buildings and the railroad line out of the floodplain. The city’s share of costs was paid for in a 0.5% rise in the sales tax for 20 years; this produced about $3.9 million a year. The Corps removed the dams and levees along the purchased stretch of river to let the river reclaim its historic floodplain. The result was an immediate rise in property values, as well as an estimated $22 million saved annually in flood damage, $4 million in the reduced cost of flood insurance, cleanup and emergency aid (some of these costs would have been borne by the federal government). In a variation on this idea, the government of Costa Rica pays farmers to restore native forests on parts of their land; the forests reduce the sediment loads in rivers, which lengthens the lifetime of reservoirs; they absorb carbon dioxide (which makes the government able to claim carbon credits), are useful for pharmceutical exploration, and attract tourists. New York City recently began to buy more of the land about its upstate reservoirs, and is trying to control land use on the rest, in order to to keep the water in the reservoirs clean enough so it can avoid building a filtration plant that would cost several billion dollars, plus yearly operating costs of perhaps a billion. If sellers were willing, buying all the land in question would be cheaper than building the plant (at the time purchase started, the private property in the watershed was worth about one billion dollars). This would have been a permanent solution, but not a politically possible one. But all settled landscapes should reduce their impact on waterways; in a very real sense, all watersheds contain drinking water.

* * *

Modern agricultural landscape are new landscapes. They haven’t existed before. Eugene Odum recommended keeping 40% of any ecosystem in its original state (let’s say, in a biologically working state; since the original state was always evolving and changing). If farm fields were tucked into a larger landscape of forests or prairie their leakiness would make much less difference. The nutrients seeping from them would be turned into trees, deer, prairie grasses, game birds, or recycled in riverbank wetlands into bacteria, plants and invertebrates; some would feed fish. But in most of the world, turning the landscape and its resources into commodities has destroyed the ability of the pre-existing system to work, biologically speaking; streams are drains; the natural environment is pulled apart; and farmers focus on their fields and crop yields alone. Our current landscape has been created by capital and government policy, with farmers as more or less wlling accomplices. A major part of farm income (half, or $28 billion, in 2000) comes from price supports for commodity crops. Modern price support policy favors unlimited production, in order to keep food prices low. The money is recycled to producers of seeds, chemicals and farm machinery, who, like consumers, have an interest in maintaining the status quo. Export enhancement programs help large grain companies and food processors market their products abroad. Shippers benefit from tax-supported highways and locks and dams on rivers. Seed companies, chemical companies, meat producers and the government sponsor research on increasing production at land grant colleges.

But policy—through programs like the Conservation Reserve Program, through looking at an agricultural landscape as a whole—can also create new landscapes. What would an agricultural landscape that produced grain, meat, habitat for native plants and animals, fishy rivers, jobs, healthy estuaries, be like? In heavily agricultural areas, 15-30% of the landscape would be uncultivated land. Its purpose would be to absorb the nutrients and slow the water running off the fields. It would also provide habitat for native wildlife, much of which is useful to the farmer. (The annual value of pollination by wild insects and of insect and rodent control by wildlife in the United States is calculated at about $50 billion.) Foxes, hawks and owls control rodents; pollinators need to be near the crops they pollinate. Pollination is most efficient when the insects have to travel less than a hundred yards, so fields that require pollination should be no more than two hundred yards wide, bordered by wide hedgerows or other later successional landscapes that will support a healthy population of pollinators. In Costa Rica coffee yields are 20% higher in plantings within a thousand yards of a national forest. Much of the protected land would be along watercourses, some of it would be hayed once a year, some lightly logged, some of it would be in constructed wetlands meant to filter and slow field drainage (some of these lands could also be hayed late in the summer). Studies in South Dakota found that leaving 15% of an agricultural area in undisturbed land (in this case, Conservation Reserve Land) gave the most economic benefit in terms of wildlife (here mostly considered as game animals); less land provided too little space for the animals, and the economic benefits per dollar fell with more wild land (the ecological benefits would have risen, however). Such natural areas should connect. Not all land is now cultivated even in the Corn Belt (though 89% of Illinois is cropland). Some of the land that would be used to protect streams is already steep and eroding and many farmers would be happy to see it protected, especially if someone else paid for the grading and revegetation. Wild landscapes function better if the agricultural landscape itself is more benign—with mixed crops, smaller fields, less use of pesticides and herbicides, more rotational grazing (more animal flesh produced from grass), fields cut or harvested at times that allow birds to finish breeding. The two then merge together.

What would this landscape produce? Land taken out of production to reduce nutrient runoff and surround farms with a working biological landscape (producing public goods) would cut the current corn crop by several percent, say 5-10%. Rotating row crops (corn, soybeans) with hay and small grains (wheat, oats, barley, which function to a certain extent as sods) cuts the corn crop by probably 30%; the small grain crop on the other hand would rise. Organic agriculture, or a limited use of nitrogen fertiliser, cuts the corn crop by another 5%. So the annual corn harvest is now down by 40-45% (from 9 billion bushels to perhaps 5 or 6: the crop a decade ago). One result of this will to raise the price of corn and make growing it profitable again. (Modern genetics has been so successful in raising corn yields that in 2006 even agricultural economists say the best thing that could happen to farmers would be for the corn crop to fall by 15%. The size of the corn crop led to federal support for the conversion of grain to ethanol, a motor fuel, in order to absorb some of that corn; and that, combined with rising demand from the developing world, has currently—2008—raised the price of corn to profitable levels. A 45% drop is a big drop, but it won’t happen all at once and further progress in plant breeding is likely to raise yields further.) Since most of the United States corn crop goes to feed animals, and since, in a regenerative agriculture, animals would be raised on hay (and some grain) produced on the farm, and since feeding cattle corn and soybeans in feedlots is tremendously wasteful of feed, as well as harmful to the animals, the agricultural productivity of the landscape (the calories it produces in grain, meat, milk) should not fall by anything like the reduction in the corn crop. It may fall by more than the amount of land taken out of cultivation (say 20% instead of 15%) Some of the land taken out of production should add to farm income through tourist dollars or hunting leases. Farm income in general under the new regime should be higher, since costs of production will be less and farm prices higher.

My figures are of course as suspect as anyone else’s. About 30% of the calories fed hogs and cattle currently come from recycled animal fats. (Another 20% of animal feed comes from the by-products of food processing, that is, from the cake left by oil seed pressing, from rice bran, peanut shells and skins, distillery mash, citrus pulp, other grain milling wastes. Cattle may also be fed the high fiber wastes of other domestic animals.) Because of mad cow disease, most countries no longer allow cows to be fed cow remains. The United States has been slow to admit the danger to its food supply and here calves are still raised on cow blood and fat. Pigs and chickens are still fed animal fats. Such practices are extremely dangerous biologically. (The ritualistic eating of human brain results in a neurological disease in humans; and eating our close relatives the chimpanzees and gorillas resulted in the adaptation of HIV/AIDs to people.) Bags of feed can get mixed up (so pigs eat pigs and cows eat cows). Moreover pesticides and other cancer-promoting and hormone-mimicking chemicals to which farm animals are exposed accumulate in their fat. So feeding recycled animal fats to animals (which are slaughtered and which we then eat) sets up a chain of bioaccumulation of these chemicals, as the fat is recycled over and over. Cutting this material out of the food chain (and composting it, digesting it for methane, or rendering it into biodiesel) would further lower agricultural productivity. Fat is also a food—one just shouldn’t eat too much of it. (Composted remains become a soil amendment, as they are on Joel Saladin’s Virginia farm. In the same way, the bodies of wild animals return to the soil.)

The question of farm yields and appropriate practices is difficult to summarize; there are too many possibilities. Modern capitalist landscapes tend to be extremely simplified (this is their fundamental problem): thousands of acres of one cultivar of one grain (corn in much of the Middle West); forests of even-aged plantations of Douglas firs or southern pines; suburbs of similar buildings as far as the eye can see. Such simplification leads to other problems. Planted forests of Douglas fir avoid the natural succession into deciduous shrubs and trees, mostly short-lived, that fix andf store the nitrogen on which the young firs grow until (150-200 years later) they reach maturity and use a new source of nitrogen fixed by lichens. The increasing size of the American corn crop and the falling price of corn syrup has led to Americans consuming 120 pounds of sugar a year, as bottles of sugared soda got bigger and bigger while the price remained the same; the increased sugar consumption has led to increases in obesity and diabetes. (While corn and its products got cheaper, fruit and vegetables got more expensive.) Extensive monocultures are the result of economies of scale. More varied smaller-scale landscapes are more productive per acre but not per man-hour, because of the labor involved. Smaller scales and more labor allow more innovative solutions. In East African corn fields planting napier grass in the row attracts stem borers away from the corn; the borers prefer the grass, but a sticky exudate from the grass traps the larvae. So spraying is not necessary. Also in Africa, planting a plant called Demodium keeps Striga, a parasitic pest of the corn plant, from growing in the field. Striga is otherwise uncontrollable. Small fields in parts of Africa are bordered with fast-growing leguminous trees. The branches of the trees are used as a green mulch to raise yields, and the stems yield building material or firewood. Lines of ponds dug along the valleys of natural drainages in monsoon India capture the monsoon rains and let them seep into the ground. This keeps water tables high and within reach of shallow wells, while modern methods of irrigation using deep wells and pumps lowers water tables. The stored water is used for irrigation and also provides some fertilisation. In the state of Tamil Nadhu 40,000 such ponds irrigate 2,500,000 acres. (Ponds like this, fed by winter runoff and ground water, are also used by nurseries and dairy farms in the northeastern United States and Canada.) In the Sahel, low stone walls (one stone high) laid across hillsides capture sheet flow, increasing the yield of sorghum and wheat by 70%. Both techniques could be used in the desert Southwest instead of river irrigation for small-scale agriculture. Capturing sheet flow with lines of stones is used already to restore dry (overgrazed) desert landscapes there.

Over the last 30 years farmers in Niger have allowed acacia trees to grow on their lands at a density of 20 to 40 an acre. Their crops were being buried by wind-blown sand so the farmers decided to protect the trees, which germinated naturally, and allow them to grow to shelter their fields. The trees have allowed this southern edge of the Sahara to become productive. They provide income from the sale of firewood. They shelter crops and prevent soil erosion. Their litterfall and roots enrich the soil. They provide fodder for animals, which leads to more animals and more manure, better crops, and more trees (whose seeds sprout from the manure). The trees are thought to have increased rainfall by 10-20% and have reversed desertification on 7 million acres and brought into cultivation 600,000 acres of cropland. While few modern farmers would allow trees in their fields, in California cattle pastures, cattle graze preferentially under blue oaks. The grass there is also more productive (15-100% more forage is produced under blue oaks than in the open). The cows also eat acorns and oak seedlings, as do rodents attracted to the seeds of annual grasses in the pasture, so some young oaks must be protected if the farmer wants to increase the oaks on his grazing land. In the United States, a field planted to a mix of corn, beans and squash yields 75% more food than one planted to corn alone, probably because the mixed planting makes more efficient use of water, light and nutrients. (Fields with mixed plantings also shed fewer nutrients.) Mechanical harvesting of the mix, except for cattle feed, however would not be possible. Similarly, growing corn, beans and cassava together in Cuba doubles yields. Rice yields in Madagascar were raised 3 to12 tonnes per hectare by transplanting the seedlings from smaller bundles (so they were handled more carefully, and more plants survived), keeping the paddies unflooded for much of the growing period (which saves water and reduces methane production), and using compost rather than fertiliser. This system has also been tried successfully in China and India. Japanese rice growers once used green fertiliser, a mulch of leaves and small branches collected from the woods, on their paddies. Several acres of woodland were required for each acre of rice. On a slightly different note, clever Middle Western farmers grow cover crops with a mix of up to ten species of grains and legumes to feed their animals; depending on the year, some of the plants do better than others; so yields are somewhat weather-resistant. And 30% of Argentina’s farms are no longer plowed. The farmers plant winter crops on their harvested fields (such as black oats) and spray glyphosate (an herbicide that, if it doesn’t leach into the soil, breaks down over several days into carbon dioxide and water) to kill competing weeds and grasses before planting in the spring. Grain yields are higher, soils lose fewer nutrients and continue to accumulate carbon. Using special planters, organic farmers grow no-till crops without herbicides.

Regenerative farmers use crops adapted to the place. (Of course, to an extent all rainfed farmers must use crops adapted to their soils and climate.) Winter crops can be grown without irrigation in the Central Valley of California but most summer crops cannot; some crops currently grown there (alfalfa, cotton, sugar beets) are very water demanding. About 80% of California’s water goes for irrigation. For rivers to function properly, no more than 25% of their flow should be withdrawn. More than 60% of the water that flows toward San Francisco Bay is withdrawn. So, in general, (but the San Joaquin is much more exploited than the Sacramento) use should be reduced by 35%. Crops with a high demand for water use 20% of the flow. Grapes are often grown in California without irrigation, though they need irrigation to become established. By shifting to more appropriate crops, California growers could reduce their water use considerably, without reducing the value of what they grow. John Muir, who grew irrigated plums at the northern edge of the valley, suggested (half seriously) that the best, most profitable use of the wildflower plains of the Central Valley would be to produce honey. Some areas are simply lucky. Apples grown on certain ridgelines in Mendocino County of northern California, 1200 feet above the ocean (originally the habitat of redwoods and Douglas fir) don’t need irrigation because of the fog coming in off the ocean. But the trees have more sun, and the apples ripen better, than in locations near the coast. The area is just south of the apple maggot line and west of the codling moths of the Central Valley; essentially the trees need no irrigation, spraying, fertilising or cultivation; they need to be pruned and harvested.

Animals are producers of milk and meat in industrial agriculture. The air in confinement barns (another economy of scale) is sharp with ammonia and hydrogen sulfide (both poisons; people who work in such places can be recognized by their coughs); cattle feedlots can be smelled miles away; the animals are tense, crowded and denied their instinctual behaviors. The main job in raising chickens in confinement barns, apart from making sure the machines continue to provide feed and water, is removing dead birds. Pigs are a traditional animal of the Corn Belt, a means of marketing the corn crop as meat. Pigs, like chickens, are much more efficient converters of grain to meat than cattle (though the conversion efficiency of milk is good). Pigs were once raised more or less outside in much of the Corn Belt; now they are raised in confinement facilities, often under corporate sponsorship (corporations provide the facility and the grain; and buy the pork; the farmer provides the land and labor). Confinement facilities have their problems of odor; outdoor manure lagoons that fail in rainstorms; fighting and cannabalism among the animals; up to 20% of the pork of too poor quality to sell as fresh meat. Some argue there is a moral cost to treating animals so badly (as I argue there is one to so badly abusing a landscape). An alternative to confinement buildings are movable fabric tunnels. The animals are bedded over a deep layer of litter (hay or straw), in which they can root and build nests. They are also free to move around and socialize. The composting of the litter layer (one to two feet deep) helps keep the animals warm in winter and eliminates much of the indoor odor of ammonia and hydrogen sulfide. Allowed to be pigs, the animals are much less stressed; costs are 60% less than in confinement barns. After the pigs in a tunnel are slaughtered (perhaps one day this will take place on the farm, in familiar surroundings, without the stress of that final journey), the tunnel is taken down and moved, the bedding scraped up in a pile to finish composting, then spread on the fields. In a world more attuned to local markets, the ground under the tunnel, cleared of sod by the animals and saturated with manure and urine would be used to grow vegetables. The tunnels are large; one can also rotate them over the fields, using the land underneath to grow corn. In the warm season the hogs are raised outside in pastures (after a season’s use, those pastures could also be planted to crops; first cover crops, then vegetables or grain).

The English landscape before World War I developed under the twin pressures of capitalistic agriculture and the land’s use for hunting. (Following World War II, the landscape became more and more one of modern commodity agriculture.) Many of the common lands of England became aristocratic lands after the 1200s, with the extinction of the hunting rights of commoners (the time of Robin Hood). Privately owned lands without heirs passed to the local noblemen, who accumulated many lands during the 1400s and 1500s, the centuries of the Black Death. During the 1500s more common lands were enclosed by large landowners, with the justification of increasing the national prosperity by raising more sheep for wool. Later, as the value of manure on cropland became clearer (one acre of grain needed the manure of half a cow), Parliamentary enclosure laws allowed rich landowners to enclose the remaining common lands for pasture for their cattle. In return tenant farmers gained heritable rights to their farms. But English estates were managed for both agriculture and game. (Hunting, once training for warfare, retained its grip on the noble mind.) Entailment of the estate to the oldest son meant long-term development of the land was possible. Labor in England was cheap. So the fields were bordered with hedgerows that bloomed with wildflowers in the spring and provided habitat for birds and animals, while rough woodlands (copses) were maintained for pheasants for the fall hunt. (Red grouse moors were in Scotland.) The meadows with their weedy edges were for Hungarian partridge, as well as for pasture and hay. Predatory birds and mammals were trapped and shot. Wildlife was abundant but not diverse, with few wide-ranging herbivores or carnivores. The owners of the estates knew what they wanted, whether one agrees with it or not, and ran their farms both to make a profit and to enjoy their landscape. Maintaining the hunt was regarded as a moral imperative. After the 1950s the profit motive took over and the old landscape started to disappear. (By contrast, in the United States farming became commercialized almost upon arrival. Speculation soon broke up the town plans of the Puritans, which allocated to each family the amount of land it could use — though the rich, who could hire labor, got more. In 1910 in the United States, half the farmers had been on their farms for less than five years, and half the land in the Corn Belt was rented.)

I once saw a drawing by a Mexican artist of his home. The small houses of the village were on a terrace above the river, among fields of corn, the river below. In the background was a forested mountain with an eagle perched on top. Agricultural landscapes like this are also new landscapes, but ones that better fit into the biologically functioning habitat of the continent, with its rivers and seas, eagles, migrating butterflies, geese, fish. Agricultural landscapes in this sense become the second nature of the garden; maintained not only for profit, but for the way they fit into a whole landscape. Farming becomes much more of an intellectual game. In this new landscape large trees in the gallery forests along rivers attract breeding ospreys; hollow trees shelter breeding owls or kestrels, Cooper’s hawks nest in dense clumps of backyard evergreens, merlins and sharp-shins ambush blue jays at the feeder: such birds also hunt mice, rats, grasshoppers, English sparrows, starlings, rock doves, and other endemics of the agricultural landscape. The absence of large carnivores (wolves, mountain lions, bears) means more abundant mid-size carnivores (skunks, coyotes, foxes, raccoons, oppossums) that eat birds and smaller animals (amphibians, Norway rats, rabbits, mice; coyotes also eat deer, especially the fawns, and in the East seem to be evolving into a larger animal, in order to be a more efficient predator on them). Cliff swallows nest on the sides of barns, barn swallows under the roof overhang, bats roost under boards under the eves. Mushrooms grow in the woods and perhaps are cultivated by the farmer. Unplowed springtime pools in the fields provide habitat for migrating ducks and shorebirds and for breeding amphibians (these are also aquifer recharge areas that filter runoff from the fields); in the Southeast, such damp, thorny tangles are habitat for quail. If grassed, aquifer recharge areas can be mown late in the season, or grazed (along with the cover crop); if wooded, lightly logged. The fields themselves grow mixtures of crops; wide hedgerows (100 to 200 feet, so-called ecological edges) break the wind, shelter helpful insects and birds, sloughs, ephemeral streams. Agricultural support payments no longer support crop prices but compensate farmers for maintaining a working biological landscape, with healthy streams. How the farmer maintains a healthy landscape is up to him. Not all biodiversity is functional, but all of it lifts the spirit. Thus one creates a new landscape