Courtesy Agricultural Service, USDA

Land-use change in the United States represents an enormous uncontrolled experiment in the ways habitat changes influence plants and animals. When cities were built, land was plowed, or forests were cut, the effects on our native biota were not considered. Of course, humans have influenced the flora and fauna of North America over the ages. Native Americans established settlements, practiced agriculture, hunted, and used fire to induce vegetation changes (Denevan 1992). Land-use changes, however, have been particularly profound since Europeans settled North America three centuries ago. Landscapes have become mosaics of natural and human-influenced patches, and once-continuous natural habitats are becoming increasingly fragmented (for example, Burgess and Sharpe 1981; Harris 1984). Our nation's lands experienced tremendous changes in response to human activities, but our understanding of how land-use history affected the plants and animals in our modern landscapes is incomplete.

The term land-use change has several meanings; we use it to include changes in both land cover and land use. Land cover refers to the habitat or vegetation type present, such as forest, agriculture, and grassland. Land-cover change describes differences in the area occupied by cover types through time. Both losses and gains are included. For example, changes in how much forest occurs across a landscape may reflect additions as croplands or rangelands are abandoned and forests regrow. Forest cover may also be lost to harvest or to development. In addition to tracking the amount of cover types, land-cover change also describes shifts in the spacing of cover types across the landscape over time. For example, forest may occur in a large block, or it may exist as several smaller parcels. Land use is usually defined more strictly and refers to the way in which, and the purposes for which, humans employ the land and its resources (Meyer 1995). For example, a place that is in forest cover may be used for low-density housing, logging, or recreation. Land-use change encompasses all those ways in which human uses of the land have varied through time.

Land-use patterns have important influences on biological diversity--that is, the abundance, variety, and genetic constitution of native animals and plants--for several reasons. First, land-use activities may alter the relative abundances of natural habitats and result in the establishment of new land-cover types. The introduction of new cover types can increase the variety of species by providing a greater diversity of habitats. Natural habitats, though, are often reduced, leaving less area available for native species. Species that are not native to the locale may gain a foothold and out-compete the native species (see chapter on Nonindigenous Species). Second, the spatial pattern of habitats may be altered, often resulting in the fragmentation of once-continuous habitat. The effects of habitat fragmentation on animals, plants, and their habitats are numerous (see summaries by Saunders et al. 1991 and Noss and Csuti 1994), and the biological diversity of native species is almost always reduced. Third, land-use activities may change the natural pattern of environmental variation, especially by causing changes in natural disturbance patterns. For example, the environment may be changed directly when fire control and logging alter the frequency and extent of natural fires. Environmental scientists understand that natural disturbances create and maintain biological diversity by creating a mosaic of habitats (see chapter on Natural Processes). In general, the chances of losing native animal and plant species and disrupting ecological functions increase when the patterns of natural habitats are altered.

The relation between land-use change and animal and plant communities sets the stage for understanding our present-day fauna and flora and for managing our lands. Both worldwide and in the United States, land cover today is altered principally by direct human use: agriculture, raising of livestock, forest harvesting, and construction (Meyer 1995). Our human society relies on natural habitats for a variety of services including productivity, recycling of nutrients, breakdown of wastes, and maintenance of clean air, water, and soil. The rate and quality of these services, and their capacities to recover from stress, often depend directly on communities of native plants and animals (Schulze and Mooney 1993; Tilman and Downing 1994; Mooney 1996). The ways in which plants and animals respond to new kinds of human-caused environmental disturbances can help us learn more about the tolerances of species and their needs. We present a historical perspective on land-use change and its effects on biological diversity in the United States and consider what this may teach us and what to anticipate from current land-use trends.



The vegetation in our nation's landscapes is the result of a combination of environmental characteristics (such as soils, climate, and topography) and land-use history (Foster 1992; White and Mladenoff 1994; Meyer 1995). At the time of European settlement, forests covered about half the present lower 48 states. Most of the forestland was in the more moist East and Northwest regions, and it had already been altered by Native American land-use practices (Williams 1989). Clearing of forests for fuel, timber, and other wood products and opening the land for crops led to a widespread loss of forest cover that lasted through the early 1900's (Fig. 1). So extensive was this loss that by 1920 the area of virgin forest remaining in the contiguous United States was only a tiny fraction of that present in 1620 (Fig. 2).

Cropland increased at the expense of other land covers throughout much of American history (Meyer 1995). Cropland reached a peak in the 1930's and has since fluctuated around 162 million hectares (Fig. 1). Irrigated cropland became significant only in the late 1800's, and it increased rapidly after 1945, primarily in the drier West, with the advent of more advanced technologies like center pivot irrigation. By 1980, about one-eighth of all cropland was irrigated. While dry cropland was being irrigated, however, wetlands were being drained for agriculture. Between the 1780's and 1980's, 53% of American wetlands were converted to other uses (Meyer 1995). Between the 1950's and 1970's alone, nearly 4.5 million hectares were lost. Grasslands also experienced a net decline (Fig. 1).

Developed land in the United States has expanded as the population has grown. Most of the population of the United States now lives in cities, towns, and suburbs rather than on farms. Americans spread out across the land as transportation technologies improved, especially as the automobile became the primary mode of transportation. Present-day patterns of settlement involve more land per person than in the past, and homes and subdivisions are more dispersed across the landscape. A frontier of rapid and sometimes chaotic land-use change surrounds urban areas (Meyer 1995). Trends in developed land are unique because they run in only one direction--that is, developed land expands and does not revert to other categories. Thus, the distribution of developed land across our nation will leave a long-lasting footprint on our landscapes.

Ninety-seven percent of the nation's surface area is land, and consequently most of the water that enters streams and lakes is affected by activities on land (National Research Council 1992). Declines in stream and river conditions in the United States are well documented (Naiman et al. 1995). Of the 1 billion hectares of land in the contiguous United States, nearly half have been cultivated or grazed by livestock (Bureau of the Census 1990). Changes in drainage and erosion that accompany agriculture have substantial effects on fresh waters (National Research Council 1992). Urban areas account for only 3% of the land in the United States, but the effects of densely settled areas on our water resources are extreme. As cities are built, natural wetlands and floodplains are replaced with surfaces such as asphalt, cement, and buildings that cannot absorb water. Flash flooding becomes common, and additions of waste and pollutants to downstream waters increase (National Research Council 1992). Thus, land-use change affects not only the land but also the waters.

The timing and sequence of land-use changes in the United States have varied among regions. Territories of the United States were acquired from east to west as land was purchased and treaties were signed (Fig. 3). Settlement patterns also followed this trend as new lands were acquired by an expanding nation and people were encouraged to move

westward (Fig. 4). Clearing of forests rapidly followed settlement, beginning in the East and proceeding westward. Some originally cleared areas--such as New England, the Southeast, and the upper Midwest--have now become reforested because of lack of cultivation (Fig. 5). In other regions, clearing for agriculture has been more permanent (for example, the lower Midwest), or harvest of virgin forest has continued until recent times (Pacific Northwest). In this section, regional patterns of land-use change are compared to illustrate the diversity of changes across the country.



The eastern United States has undergone dramatic changes in the area covered by forest during the past 250 years. Foster's (1992) analysis of land-use change in Petersham Township, central Massachusetts, illustrates the cycle of forest clearing and regrowth that was typical of much of the Northeast. Forest clearing by European settlers began around 1730 and proceeded at a pace of 1%­4% of the original forest per year. Open land covered half of Petersham Township in 1800, and nearly 85% of the land was open by 1850 (Figs. 5 and 6). Remaining uncleared forests generally occurred in less accessible locations, such as steep slopes, but were still used for grazing, timber, and fuelwood. Farming became unprofitable and farm abandonment commenced in the mid-1800's; thereafter, forests gradually grew back on neglected fields. This process led to the broad-scale establishment of forest cover over the entire region. Most recently (1920­1990), Petersham Township is characterized by residential development and almost no agriculture.

The patterns described by Foster (1992) were generally repeated throughout the northeastern United States. Ninety-five percent of Dutchess County, New York, was in farmland as late as 1880, but farming had declined dramatically by the 1930's (Glitzenstein et al. 1990). Indeed, much of New York State followed a similar pattern (Hill 1985; Nyland et al. 1986; Smith et al. 1993). Forest regrowth, however, was not always the rule. Forest cover in Franklin Township, New Jersey, increased between 1880 and 1943 but then declined; only 16% of the area was forested in 1986 (White et al. 1990).

The oak­hickory forests that range from the Ozark Mountains to the Appalachian Mountains owe their current condition largely to human activity, which began with Native American activity before 1600 (DeVivo 1990; Denevan 1992). Native peoples used fire to move game animals, to open up the forest, and to clear underlying brush and saplings. The native peoples also cleared land for agriculture in river valleys and on flat, moist upland sites (Parker 1991). Their use of fire and forest clearing affected the types of trees found in the forest communities, so that the widespread dominance of oak and hickory most likely had its origin in Native American practices.

The European settlers who began to arrive in the Midwest in the late 1700's initially adopted Native American methods of clearing forests. Much of the land cleared in the early 1800's was abandoned as soil fertility declined and settlers moved farther west. But as technology developed by the mid-1800's, forests were permanently cleared. In Illinois, for example, only 31% of the forest area present in 1820 exists today (Iverson 1991; Fig. 7). Until 1860 forests were the sole source of potential cropland, and by 1860 the timber industry was thriving in Illinois. Iverson (1991) makes a striking comparison of deforestation rates in Illinois from 1820 to 1923 and recent rates of forest clearing in three tropical countries (Table 1). It is instructive to note that deforestation in the United States was of a similar magnitude as present-day deforestation rates in the tropics.

The central hardwood forest ecosystem of the Midwest is diverse and resilient. The region apparently has survived massive human disturbance with the disappearance of few native species. Some species are rare because their habitat has declined or because they were harvested by humans. Many other species are rare because of a reduction in environmental patchiness (see chapter on Natural Processes) due to natural disturbances such as periodic fires and storms (Anderson and Schwegman 1991).

Along the boundary between forests and prairies, it was the forested lands that were first settled. Settlers believed that the prairie was unsuitable for settlement because of lack of wood and drinking water, lack of protection from winter storms, and the lack of trees, which was thought to be caused by infertile soils (Latta 1938; Anderson 1970). Settlement began in the early 1800's, but even as late as 1836 only small prairies near woodlands were settled, despite the invention of plows capable of tilling the thick prairie sod and the high crop yields obtained from prairie soils (Anderson 1970). After the railroads were established in the 1850's and 1860's, crops and supplies could be transported economically, and the prairies were then rapidly settled (Barrows 1910; Anderson 1970).

Cattle raising was largely a frontier activity, and by 1865 the cattle industry had moved to Illinois, Iowa, and Missouri and was pushing farther west. Spanish cattlemen introduced herds to both Texas and California before 1865 (Rasmussen 1974). Herds of cows roamed freely on the plains with little tending and by the 1880's had expanded throughout the northern plains (Fedkiw 1989). After World War I, ranch farming became more common because it reduced some of the economic hazards of herding. Cattle numbers on the plains have varied widely in this century because of unpredictable markets and weather. Range conditions were probably at their worst in the late 1920's.

The decline in grasslands (Fig. 1) was primarily due to a shift to cropping, mostly wheat. Beginning in the 1880's, dryland farming and irrigation development in the semiarid West were encouraged by the government and by the railroad companies. By 1920 more than 75 million hectares had been added to annual cropland, mostly west of the Mississippi (Fedkiw 1989). Today, very little of the original extensive native prairies remains.

Cutting of forests in the Pacific Northwest began in the 1800's when European settlers arrived in the region. The extensive forests covering the lowlands and river valleys were viewed as blocks to progress and were systematically burned and cleared for agriculture (Forest Ecosystem Management Assessment Team 1993). In the late 1800's and early 1900's, commercial logging began to increase on the uplands but did not begin in earnest on federal lands until after World War II. European methods of forest management were gradually adopted on most federal and private lands--techniques such as clear-cutting, removal of logs and snags, slash burning, thinning, and planting stands of a single species. The assumption was that forests managed in this manner could be cut and regrown at relatively short intervals (for example, 40­80 years) without negatively affecting water quality, fish, soils, or land animals. Stream and riparian ecosystems in the region first suffered damage from grazing and mining that occurred from the 1860's to 1910, and then from the logging, roads, fire management, and irrigation that occurred in the twentieth century (Wissmar et al. 1994).

As a result of more than a century of logging and fire control, the forests of the Pacific Northwest presently consist of a highly fragmented mosaic of clear-cuts, thinned stands, and young (and often single species) plantations interspersed with uncut natural stands. In addition, the development of extensive road systems to access the forests has further modified the landscape. The road density often equals or exceeds the density of natural stream channels in the basin--more than 2.5 kilometers of road per square kilometer. Only about 15% of the 10 million hectares dominated by Douglas-fir remains as ancient forests in the 12 national forests studied by Morrison et al. (1991); the percentage appears to be less on private and state lands. The remaining ancient forest, for the most part, occurs in highly fragmented stands and is subjected to greater damage from windstorms and human intrusions caused by edge effects and the extensive road system. Many of the remaining stands are in small patches (less than 30 hectares) and within 125 meters of roads or clear-cuts, significantly decreasing their value as habitat for native plants and animals.

The Olympic Peninsula in Washington state provides a regional illustration of the changes that have taken place throughout the Pacific Northwest (Fig. 8). Since 1940 nearly 76% of the Olympic Peninsula's ancient forests have been logged (Fig. 8a). In 1940, 87% of the ancient forest was in patches greater than 4,000 hectares, whereas by 1988, the forest was so fragmented that 60% of the ancient forest occurred in patches of fewer than 40 hectares (Fig. 8b). In addition, very little of the low-elevation forest remains today. In the Pacific Northwest, many species are only found below elevations of 1,200 meters and others are only found below 600 meters. Most of this species-rich lowland forest occurs in small stands or fragmented webs with the natural communities endangered by their proximity to forest edge (Fig. 8c).

In the late 1870's, following serious efforts by European immigrants to subdue the Native Americans, much of the Southwest was opened to settlement (Cooper 1960). The history of settlement and vegetation change of southwestern pine forests since settlement has been documented by Cooper (1960). Early accounts of these forests stressed the open nature of the woods, luxurious grass swards, and lack of gullying. It is likely that the open nature of the forest was maintained, in part, by use of fire by Native Americans. Some tribes, such as the Navajo, also maintained extensive herds of stock. The introduction and subsequent mismanagement of livestock in the West, however, produced profound changes in plant cover (Cooper 1960). Deterioration of watershed condition was rapid following settlement, reaching its worst state by the turn of the century and leading to serious concern over water supply. The adverse ecological effects of overgrazing are well known (for example, see West 1993).

Cumulative effects of land-use changes in the Southwest since settlement include a decrease in grass cover, a shift in composition of herbaceous vegetation, and an increased density of pine trees. Two factors appear primarily responsible for these changes (Cooper 1960). First, heavy grazing led to a reduction of native bunchgrasses and, in some cases, replacement by nonindigenous annual plants. The second factor was the exclusion of fire from the forests and grasslands. Frequent surface fires that maintained an open understory in the pine forests were typical before settlement by those of European ancestry. More intense crown fires were rare during this time.

In the northern Rocky Mountains, the post-1860 rush of miners greatly accelerated land-use change. Madison Valley, Montana, offers a good example of settlement patterns (Wyckoff and Hansen 1991). Although Native Americans had migrated through the area, they had not maintained permanent settlements in the valley. Permanent settlement of the region by European immigrants began around 1870; these settlers introduced large numbers of livestock to the land, resulting in deleterious effects on many habitats. As range conditions deteriorated, fescue and wheatgrasses declined and were replaced by rabbitbrush, grama grasses, and sagebrush. Seasonal overgrazing by livestock caused particular damage to sensitive riparian and subalpine habitats (Wyckoff and Hansen 1991).

Recent land-use changes in the West are dominated by increased demand for housing development. In California, for example, the human population increased 25% between 1980 and 1990, leading to urbanization of 123,000 hectares of previously undeveloped land (Charbonneau and Kondolf 1993). Smaller cities in the Rocky Mountain states are also experiencing intense developmental pressure. Residential development is occurring nearer and nearer to large private and public landholdings; this geographic proximity makes forest and wildland management problematic. For example, management of natural fire becomes more of an issue when homes and businesses are nearby. In addition, dispersed residential development leads to extensive natural habitats becoming increasingly fragmented at fine scales. The sizes and connectivity of native habitats become reduced, making wildlife management more difficult.

The composition of plant and animal communities across the United States has undoubtedly changed continuously over the past 400 years as human uses of the land have changed. When Europeans arrived in North America, there were approximately 1.9 million people occupying the vast continent (Ubelaker 1988 in MacLeish 1994). The level of human activity on the land diminished rapidly in the sixteenth century as the large Native American populations dwindled following the introduction of European diseases (Denevan 1992). Human activity then increased rapidly as Europeans settled the landscape in the seventeenth and eighteenth centuries (Whitney 1994).

Although these general trends are conspicuous, remarkably few studies have addressed the changes in species abundance or composition associated with land-use change since Europeans came to America. The kinds and proportions of trees in forests can sometimes be reconstructed from historical records such as General Land Office surveys and compared with the makeup of present-day forests. Although past land use can be compiled from deed and tax documents, there is little numerical information on the abundance and distribution of most nonwoody plants and most animals. Birds are the best-studied animals, yet most information on them does not extend back farther than the 1950's.

Are present-day forests similar in composition to the forests encountered by the early settlers? Dramatic changes in forest cover have occurred throughout the eastern United States and upper Midwest during the past two centuries, suggesting that our forests displayed considerable resilience--that is, the ability to recover from disturbance. On a regional basis, the distribution of forest types closely resembles the presettlement patterns, but on closer inspection, there are substantial differences in present-day forests.

In the 1700's in Petersham Township, Massachusetts, regional forest communities were distributed along slopes and valleys much as they are today (Foster 1992). Trees such as birch, red maple, and oaks, which are found in young forests or which can resprout, have increased in number, however. At the same time, there has been a decline in long-lived, shade-tolerant species such as eastern hemlock and sugar maple. At a local scale within a forest, the effects of land use are dominant. For example, the variety and abundance of different trees vary substantially among areas that were permanent woodlots, pastured, or cropped (Foster 1992).

When species composition is examined, even patches of forest that seem mature and stable may be very unlike the forests that existed before the arrival of European settlers (Foster et al. 1992). A single woodland that remained forested throughout the settlement period has changed profoundly in tree species composition and structure during the past 250 years. The dominant species in the canopy have changed at least three times. Surprisingly, many of the trees that were common before European settlement are now rare within the woodland and uncommon in the landscape. Sugar maple and American beech have been almost completely eliminated locally because of land-use changes (Foster et al. 1992).

According to detailed vegetation and land-use analyses (White and Mladenoff 1994), similar transitions in forest composition have taken place in northern Wisconsin. Old-growth eastern hemlock and mature hardwood dominated a 9,600-hectare study landscape during the mid-1800's. By 1931, after large-scale logging and burning of the slash left behind, young forests covered more than 50% of the landscape. By 1989, the area was covered by a mixture of second-growth hardwood and conifer types. Regionally, the regrowing young forest closely resembles presettlement forest distribution, but the variety of trees has changed. A return to the formerly dominant hemlock forest in northern Wisconsin is not likely, based on current trends. Rather, there are two likely alternative directions leading toward either northern hardwoods or a boreal forest dominated by spruce, fir, and pines. Other midwestern oak forests are being converted to forests dominated by maples, black cherry, and yellow-poplar. This trend of reduced oak regeneration has also been documented in Pennsylvania (Abrams and Nowacki 1992) and Illinois (Iverson and Schwartz 1994). Thus, modern forests have not returned to their presettlement composition (Foster et al. 1992). Although the total area covered by forest has increased since 1900 across much of the eastern United States, today's forests are very different from those of the 1700's.

Animal communities have been profoundly affected by the removal and regrowth of forests--we know that animals inhabited places that would surprise us today (see MacLeish 1994). Bison used habitats ranging from the Great Plains all the way to Florida, grizzly bears ranged the Midwest, and elk roamed New England. Passenger pigeons, now extinct, flew in great flocks. Qualitative changes in the presence and abundance of animals during the past few centuries have been described. For example, predators such as wolves and furbearers such as beaver declined dramatically following settlement. French and Dutch fur traders in the Northeast took 30,000 beaver pelts in 1620 and almost 300,000 in 1690 (MacLeish 1994), but few such quantitative data exist. Although species have been lost and gained over past centuries, detailed studies have been conducted only in the past several decades.

Forest removal and regrowth create a series of young forest stages that change in space and time. Much of the young successional forest that dominated the landscape earlier during this century has declined markedly as these secondary forests approach maturity (Brooks and Birch 1988; Powell et al. 1993). Animals that thrive in early successional habitats also increased in abundance early in the century and are now declining. For example, as forest cover across the New England states increased to 75%­90% during the past century (Irland 1982), the New England cottontail declined substantially throughout its range (Chapman and Stauffer 1981); the pattern of decline correlated with losses of old fields and young forest habitats (Litvaitis 1993). Bobcats in New England experienced a simultaneous decline related to the decline of their early successional prey species (Litvaitis 1993).

Changes in bird abundances correlate with changes in early successional and forest cover. During the 1937­1989 period of reforestation in Massachusetts, 8 of 18 migratory bird species that declined are known to nest in early successional habitats (Hill and Hagan 1991). Since 1966 the eastern towhee--a species that prefers shrub cover, including abandoned agricultural fields--has declined 13% in the eastern United States (Hagan 1993). Between 1966 and 1988 in coastal New England, 9 of 18 species of nesting land birds associated with early successional habitats have declined (Witham and Hunter 1992). Agricultural land decreased by 9% and early successional habitats by 12% during the same period, while rural housing area increased 23% and urban­industrial land use increased 4%. Additionally, many New England breeding birds that winter in the Southeast have declined significantly in recent years. It is possible that land-use change in the Southeast may be an important factor contributing to this decline as well (Hagan et al. 1991).

In contrast, forest birds may be rebounding as open habitats diminish. Nine of 15 species in coastal New England that have increased in numbers since 1966 were associated with closed canopy forests (Witham and Hunter 1992). From 1953 to 1976, changes in the bird community of a Connecticut forest preserve were correlated with changes in the abundance of forest in the surrounding landscape (Askins and Philbrick 1987). Forest birds in the preserve declined with forest removal in the surrounding landscape but began to increase again when forest cover increased. Bird species that thrive in residential areas also increased as the surrounding suburb expanded.

Although forested habitat has increased in some areas, today's forested habitats may not be as suitable for some native species when compared to the forested habitats of presettlement times. For example, the extensive deciduous forest that covered the southeastern Piedmont has been replaced by pine forests, which do not provide the conditions needed by various species--forest understory wildflowers and birds such as the red-eyed vireo--that thrived in the original deciduous forests. In the Midwest, old-growth forests and disturbed forests differ substantially in their spatial pattern across the landscape (Mladenoff et al. 1993). Compared with the large, continuous old-growth forests, the disturbed forests have numerous smaller patches of simple shape. Furthermore, certain habitat combinations--such as eastern hemlock forests aligned next to other conifer forests--occurred in the old-growth forest but are absent in the disturbed forests. Such habitat alignments are important for species that use edges or obtain resources from different habitat types (Dunning et al. 1992).



Wetlands, floodplains, and riparian vegetation zones have often been altered by agricultural and urban development (Fig. 9). Woody riparian vegetation once covered an estimated 30­40 million hectares in the contiguous United States (Swift 1974); at least two-thirds of that area has been converted to nonforest land uses, and only 10­14 million hectares remained in the early 1970's. Floodplain clearing for agriculture, urbanization, and water-resource development (especially channel modification and impoundment; see chapter on Water Use) is responsible for much of the loss of riparian forests. In many states of the arid West, the Midwest, and the Lower Mississippi River valley, riparian forests have been reduced by more than 80% (Swift 1974).

Sedell and Froggatt (1984) provide a classic example of the loss of riparian forest for the Willamette River, Oregon. Before 1850 the streamside forest extended up to 3 kilometers on either side of a river characterized by multiple channels, sloughs, and backwaters. By 1967 government-sponsored programs for forest clearing, snag removal, and channelization (channel deepening and straightening) reduced the Willamette River to a single uniform channel that had lost more than 80% of its forest and land-water edge habitats. Fresh waters are especially sensitive to changes in these adjacent lands (Osborne and Kovacic 1993). About 30% of the wetlands of the 50 states and 53% of the wetlands of the contiguous 48 states have been lost (Dahl 1990). The former wetlands slowed the flow of water, and their loss greatly increases the chances and severity of flooding. Wetland conversion also leads to loss of plant, wildlife, and fish species.

Fresh waters are degraded by increasing inputs of silt, nutrients, and pollutants from agriculture, forest harvest, and cities (Carpenter et al. 1996). Lakes and reservoirs fill more rapidly with mud from these sources. Moreover, the growth of nuisance plants, including toxic blue-green algae, is promoted by the increased silt and nutrients, a process called eutrophication. Within a lake or reservoir, eutrophication is accompanied by a loss of desirable plants as the water becomes increasingly murky. Loss of native plants and chemical changes in the water lead to loss of animal species, including fishes and waterfowl. Regional changes in land use cause widespread eutrophication of many lakes on the landscape, thereby making the lakes more similar because they are all dominated by the same set of species that can tolerate eutrophic conditions. Thus, the diversity of lake types within a landscape is reduced as all lakes become eutrophic and harbor similar species (Carpenter et al. 1996).

Lake Mendota, Wisconsin, illustrates the losses of aquatic species that accompany land-use change in watersheds (Figs. 10­14). Before the advent of the plow in the 1830's, the watershed vegetation (Fig. 10) of this lake was prairie, oak savanna, and forest (Curtis 1959). Studies of dated layers of mud from sediment cores from the bottom of the lake show that the water quality of the lake was high before the rich prairie soils of the lake's drainage were plowed (Brock 1985; Hurley et al. 1992; Kitchell and Sanford 1992). By the 1870's the conversion of native vegetation to agriculture was essentially complete (Fig. 11), and towns were growing around the lake. Large blooms of annoying blue-green algae were common by the 1880's (Brock 1985), and dramatic changes in Lake Mendota's food web also occurred as the native plant communities were replaced by agriculture (Kitchell and Carpenter 1993). The most severe declines in water quality occurred after World War II (Lathrop 1992a) when the use of agricultural fertilizers increased substantially (Fig. 12). Cities also expanded to fill about 9% of the watershed (Fig. 13), and increasing amounts of sewage were added to the lake, intensifying eutrophication (Lathrop 1992a).

Although losses of aquatic animals and plants have many interacting causes, most of the species losses from Lake Mendota occurred at the same time as did severe eutrophication, which was caused by increased cultivation, crop fertilization, and expansion of cities (Figs. 11-14). In 1946 the diversity of higher aquatic plants had probably changed little from presettlement times (Nichols et al. 1992), but by 1989, about half the species of aquatic plants were gone. The beds of wildcelery that once supported canvasbacks and other migratory waterfowl were lost, and the native pondweeds crucial for fish reproduction had also disappeared. These beneficial native plants were largely replaced by coontail and by Eurasian watermilfoil, a nonindigenous species, both of which have low food value for fishes and wildlife. Beginning around 1950, deepwater insect populations of Lake Mendota began a steep decline (Lathrop 1992b), and insect larvae - especially midges - that formerly supported fish production were almost absent from the deep waters of the lake by the mid-1960's. The long fingernailclam also disappeared. By the 1920's, six fish species had been eliminated because of intensive fishing and habitat loss. An additional five fish species

were lost by 1989 because of the disappearance of native aquatic plants and the stocking of predatory fish. Although 11 of the 37 fish species originally present in Lake Mendota have been extirpated, the lake now contains 37 fish species - the extirpated native species have been replaced by the introductions of fishes not native to the lake (Magnuson and Lathrop 1992).

Loss of native plants and animals is widespread in our nation's lakes, rivers, and other waters. The threat of a diminished variety of aquatic animals and plants is more serious than that for the variety of land animals and plants or even for the variety of organisms in tropical rain forests (Naiman et al. 1995). From 11% to 15% of the birds, mammals, and reptiles in the United States are classified as rare to extinct, but the proportion of aquatic animals and plants similarly classified is much higher - 34% for fishes, 65% for crayfishes, and 75% for pearlymussels (Master 1990). Of 214 stocks of Pacific salmon, 74% have a high or moderate risk of extinction (Nehlsen et al. 1991), primarily due to habitat loss from dams, logging, roads, and grazing (Forest Ecosystem Management Assessment Team 1993). Almost half (44%) of our nation's native clams and mussels are either extinct or endangered (Bogan 1993). Land-use-induced impoundment and inundation of riffle habitat in major river systems (such as the Ohio River, Tennessee River, Cumberland River, and Mobile Bay) and the disappearance of host fish species are responsible for the massive loss of native clams and mussels. About 82% of fishes in the United States are adversely affected by poor water quality, often resulting from habitat loss and fragmentation (Miller et al. 1989). Secondary causes of fish decline are also important, including invasions of nonindigenous species, effects of pollutants, and overfishing. Of the 27 species and 13 subspecies of fresh-water fishes that have become extinct in North America during the last century, habitat degradation contributed to at least 73% of these extinctions. The losses of aquatic plants and animals may not be apparent to the casual observer (Noss et al. 1995), but present-day lakes, rivers, and streams are substantially different from those encountered by the early settlers.

Past, present, and future land-use patterns will continue as a dominant influence on biological diversity in the United States. What general themes emerge from our understanding of past land-use changes, their effects on plants and animals, and their importance for understanding the effects of current land-use patterns? We highlight four related issues regarding land use and biological diversity that will continue to be important in coming decades: pervasive effects of habitat fragmentation, indirect effects of land-use change, the importance of landscape context, and land-use practices.

Continued habitat fragmentation is one of the most important results of recent land-use changes. Agricultural land conversion has indeed slowed during recent decades, although suburban development, which accelerated nationwide after World War II, still continues unabated. Urban and suburban developments have resulted in loss and fragmentation of natural habitats in many regions of the United States. Resource extractions such as logging, mining, and oil drilling have also resulted in habitat fragmentation.

Adverse effects of fragmentation on biological diversity have been documented by many studies (see Saunders et al. 1991; Noss and Csuti 1994). Of course, species may be directly eliminated from portions of the landscape where habitat has been converted to other uses, but the size of remaining fragments of habitat influences the number of species that can live within each fragment. Small forest patches in Maryland had fewer nesting bird species than larger patches (Whitcomb et al. 1981; Lynch and Whigham 1984); furthermore, species that nested in interior regions of a forest (that is, away from the forest edge) were unlikely to inhabit forest patches smaller than a certain size (Whitcomb et al. 1981). Most species lost from small patches were insect-eating birds that migrate to tropical regions of North and Central America. Individual birds that inhabit small forest patches also may have difficulty finding mates (Villard et al. 1993) because fewer birds are present.

The size and arrangement of habitat fragments also affect how animals use the space around them. For example, red squirrels in forest fragments had reduced home ranges in small forest patches (Wauters et al. 1994); the squirrels' use of forest resources was influenced by the size and shape of patches and the presence of connecting corridors among patches. Negative effects of fragmentation have been recorded in midwestern forests (Gottfried 1979) and grasslands (Glenn and Collins 1992; Herkert 1994), the Rocky Mountains (Keller and Anderson 1992), and western forests (Harris 1984).

Loss of old-growth forest is another noteworthy case of habitat fragmentation. Little old-growth forest remains in the lower United States, and species that require the remaining rare old-growth forests will continue to be at risk as land use continues to change. There appear to be no species dependent on old-growth forests in the East today, but old-growth obligates probably disappeared long ago and may have included the ivory-billed woodpecker and Bachman's warbler. The apparent dependency of the northern spotted owl and marbled murrelet on the old-growth forests of the West Coast has ignited an intense controversy over timber harvesting in western forests (Simberloff 1987). The concern centers on both the absolute loss of old-growth habitat and how disconnected the remaining habitat patches have become. Because fragmentation of habitat makes dispersal more difficult and reduces population stability (Lande 1988; Doak 1989), the pattern of timber harvest activity has important consequences (Franklin and Forman 1987). A common practice on national forests in the Pacific Northwest had been to maximize spacing of clear-cuts, leading to a highly fragmented forest condition that persists for many years (Wallin et al. 1994). Consequently, species that require large areas of old-growth or interior forest are thought to be at risk because of these logging practices, particularly the widespread spacing of clear-cuts throughout the forest.

Land-use activities also introduce new human-dominated habitat types such as roads, which may disconnect natural habitats. Road length and area increase with land development, and roads can be effective barriers to the movements of animals, especially invertebrates and smaller vertebrates. Mader (1984) demonstrated the almost complete division of populations of a forest-dwelling ground beetle by a road and parking loops. Of 387 small rodents that were tagged and recaptured during an extensive study on the effects of a highway in the Mojave Desert, only one was found on the opposite side of the road (Garland and Bradley 1984). Large animals are also susceptible to road effects. Grizzly bears in the northern Rocky Mountains used areas within 100 meters of roads less frequently than areas away from roads; this resulted in a nearly 9% loss of prime bear habitat (McLellan and Shackleton 1988). Roads also increased the bears' vulnerability to poachers and legal hunters. In northern Wisconsin, wolves were much more likely to use portions of the landscape where road density was lowest (Mladenoff et al. 1995). In Florida, the highway system presents both a significant barrier to movement and a cause of death for the endangered Florida panther. Even when behavioral responses (such as avoidance) to roads have been observed, the long-term effects of roads on population size and population growth rates are not known and require further study.

Many questions remain about the effects of habitat fragmentation, especially because habitat fragmentation affects different species in different ways (Robinson et al. 1992). In a study of an arthropod (joint-legged animals such as crabs, spiders, and insects) community, an amphipod population declined severely in small habitat patches, but habitat size had no effect on a scorpion (Margules et al. 1994). Some authors (for example, Haila et al. 1993) have argued that the lower numbers of species observed in small habitat patches are simply due to chance. For example, McCoy and Mushinsky (1994) found more vertebrate species in large patches of Florida scrub habitat, but when the effect of patch area was removed from the data, the small patches did not have significantly fewer species.

The fragmentation - or connectivity - of a habitat depends on both its abundance and spatial arrangement. Land-use changes generally alter both the area and configuration of habitats. If habitats decrease in area, effects on plants and animals are aggravated if the habitat also becomes fragmented. Fragmentation can result in the loss of species in single habitat patches as well as loss from the regional landscape. Land-use planning can minimize the effects of habitat loss by maintaining large blocks of native habitat and protecting natural corridors - such as river forest corridors and ridgetops - that connect the remaining large habitat blocks.

Land-use patterns and habitat fragmentation can initiate a chain of events that leads to unexpected effects. In some instances, the relative abundances of different species change in response to land use, and this in turn causes other organisms to respond. These occurrences can be considered as indirect effects because the land-use activity starts a series of events in which the final result may be unexpected. For example, elevated densities of organisms that thrive in fragmented landscapes have triggered chain reactions that affect many other species. The large numbers of many herbivores (animals that eat plants) that presently occupy our landscapes are a consequence of past land use as well as the government-sponsored predator removals of the early 1900's. Declines of species that play important roles in the ecosystem can also start a chain reaction that affects other species. For example, land-use changes on the Coastal Plain have resulted in loss of habitat for gopher tortoises. Decline of this species may cause reductions or losses of many other species because nearly 400 species of invertebrates and vertebrates have been found in gopher tortoise burrows (Noss et al. 1995).

Densities of white-tailed deer have reached unprecedented levels in many parts of the Northeast (McCabe and McCabe 1984; Witmer and deCalesta 1992; Parker and Van Kley 1993; Correll 1994). These high levels are largely due to the widespread increases in early successional habitats, the added edge habitat between forested and nonforested lands, and the reduced predation and hunting since the 1930's. Land-use changes in the past century provided bonus habitat for deer, but their overabundance has led to deleterious consequences for many other species. The high numbers of deer first influence vegetation structure and composition because the larger deer population consumes more food. Browsing damage by deer can result in significant loss of species and abundance of woody and herbaceous plants (Kroll et al. 1986; Alverson et al. 1988; Tilghman 1989; Miller et al. 1992). Deer damage can become so severe in preserves where hunting is prohibited (for example, state parks and research forests) that seedlings of trees, shrubs, and woody vines are almost completely eliminated. In preserves with dense deer populations, wildflower and grassy cover may be less than 65% of that outside the preserves (Parker and Van Kley 1993; Correll 1994). Because the number of species and abundance of forest songbirds depend on woody vegetation (MacArthur and MacArthur 1961; Karr and Roth 1971), songbird populations are depressed by high deer numbers (Casey and Hein 1983; McShea and Rappole 1992; deCalesta 1994). Clearly, land-use patterns play an important role in this story. Reduction of forest edge to reduce deer densities has been proposed (Alverson et al. 1988), but any successful solution must likely include reduction of deer numbers by hunting (Alverson et al. 1988; deCalesta 1994).

When forests are fragmented, forest birds experience higher rates of parasitism by brown-headed cowbirds - another species that thrives and increases with forest clearing. Cowbirds lay their eggs in the nests of other bird species, and the young cowbirds effectively commandeer the food and attention of the host parents, usually at the expense of the host's own offspring. Cowbirds feed almost exclusively in open shortgrass habitats, so the frequency of parasitism is primarily influenced by the amount and type of open land near the nest site (Brittingham and Temple 1983; Robinson 1992). Cowbird nest parasitism is generally limited to areas within 7 kilometers of feeding sites (Rothstein et al. 1984), although average commuting distances are usually much less (Robinson et al. 1993; Thompson 1994). Parasitism rates are generally higher near forest edges (Brittingham and Temple 1983; Yahner and Scott 1988). Nationwide, cowbirds became rarer away from the center of their range in the Great Plains and also declined as the proportion of a landscape in forest increased (Thompson et al. 1996). In the Midwest, cowbird abundance and parasitism were lower in areas with more total forest, more forest interior, and larger forest patches (Thompson et al. 1996).

Almost by definition, indirect effects of land-use change can be difficult to anticipate. These effects are passed along through the intricate web of interactions among species and may take many years to appear. It is important to recognize, though, that there may be significant and surprising long-term consequences from inducing large changes in the abundance of species.

The animals and plants contained within habitat patches are affected by what is around them - that is, the context of the surrounding landscape (Askins and Philbrick 1987; Pearson 1993). Better understanding of this broad-scale influence is sorely needed, and it must be considered in land-use decisions. Landscape context is important because the biological diversity of parks, preserves, and wildlife corridors may be affected by changes in the regional landscape (Franklin 1993). For example, the ability of a species to disperse to a particular habitat patch will depend on whether the surrounding landscape makes movement easier or more difficult (Taylor et al. 1993). Harsh and unsuitable habitats in the landscape may present barriers to species movements. River forests are often protected to maintain water quality and to provide wildlife habitat, but the effectiveness of river forest corridors is influenced by the surrounding upland landscape. For example, the species of birds found in river forest corridors are different depending on whether the surrounding landscape is forested or agricultural (Croonquist and Brooks 1993) or whether it is in the country or the city (Smith and Schaefer 1992). Urbanization of the surrounding landscape also affects other natural habitats, such as woodlots (Whitcomb et al. 1981; Tilghman 1987). The effect of cowbirds in forests, for example, depends on the presence of open feeding habitats in the broader landscape (for example, Gustafson and Crow 1994). Thus, interpretation or projection of trends in terrestrial species must consider the broader landscape.

The importance of the landscape for aquatic systems is paramount. Freshwater quality and biological diversity depend on riverine forests and floodplains that slow the transport of water, retain silt and nutrients, and provide spawning sites, food, and shelter for fishes and wildlife. Diverse, self-sustaining river and lake-edge vegetation is crucial to the quality of lakes and streams. Logging, cultivation, or residential development of river- or lake-edge forests accelerate water flow, increase soil erosion, and eliminate habitat. Such changes in these forests thereby result in significant costs in wildlife, fishes, and downstream water quality. Once the forests at the edges of rivers or lakes are significantly changed, restoration may be very difficult and extremely expensive. Thus, decisions about river- and lake-edge land use should consider costs to wildlife, fishes, water quality, and the potential permanent problems caused by development. Wetland, floodplain, and riparian restorations, which are sorely needed in many areas of the United States (National Research Council 1992), should be ranked by comparing the costs and benefits of each project. Benefits need to accumulate at the level of watersheds or even larger spatial units, and this scale must be considered in the accounting process.

Actual land-use practices, of course, affect biological diversity. Forest-cutting methods and fire reduction serve here as examples that illustrate this occurrence. However, the ways developments are sited and built, how lands are used for grazing, and the ways in which agriculture is practiced are other ways in which land-use practices are important.

Several different logging methods are widely used in North America; these have fairly predictable consequences on the structure and pattern of habitats. Even-age methods result in stands composed of trees of a single age class, while uneven-age systems periodically remove trees to maintain a variety of different ages of trees within the stand (Thompson et al. 1993). The effects of these logging methods on forest birds have been studied in ecosystems throughout the country (Conner and Adkisson 1975; Crawford et al. 1981; Steffen 1985; Thompson and Fritzell 1990; Dickson et al. 1993; Hutto et al. 1993; Thompson 1993). The common finding of all these studies is that, in the short term, logging enhances habitat for some species and degrades habitat for others. Openings produced by even-age logging upset the continuity of closed canopy forest and may have negative effects on birds of the forest interior (Porneluzi et al. 1993). Some logging methods produce so many small, dispersed openings that interior forest conditions are very difficult to maintain (Gustafson and Crow 1994). The edges produced by even-age harvests and group selection may increase predation and brood parasitism rates so much that bird reproduction is seriously reduced (Paton 1994). Uneven-age methods do allow retention of mature forest bird communities, though at lower densities than in unmanaged forests. Nevertheless, uneven-age harvest methods do not produce the same level of landscape diversity of old and young forest patches that characterizes natural forests (Dickson et al. 1993; Thompson et al. 1993).

The short-term effects of logging on bird populations have been the subject of many studies, but long-term consequences are not as well known. For example, clear-cutting may cause profound short-term changes in forests, but these changes may be critical to a long-term approach for maintaining the kind of historical ecological conditions dependent on widespread, frequent disturbances such as fire (Hunter 1992; Hutto et al. 1993; Reice 1994). The long-term effects of excluding large-scale natural disturbances from ecosystems are not well known because such protection has been relatively recent.

Concern has recently been raised about the effects of logging on land-dwelling salamanders (Petranka et al. 1993) and nonwoody plants of the forest understory (Duffy and Meier 1992). There is no general agreement on the effects of land-use practices on plants and animals that have not been well-studied; these species deserve continued investigation (Elliott and Loftis 1993; Johnson et al. 1993).

Sometimes land-use practices alter the natural disturbance regimes that generate the complex patterns of habitats required by native plants and animals. If land-use practices change the frequency, size, and intensity of natural disturbances, then altered sequences of vegetation development may lead to completely different plant communities (see chapter on Natural Processes for a complete discussion of disturbance regimes). For example, although eastern oak-hickory forests were traditionally thought of as a stable climax forest community (Weaver and Clements 1938; Braun 1950), they are now being succeeded by moist forests dominated by hard maples (Schmelz et al. 1975; Lorimer 1985; Schlesinger 1989). This trend has been attributed to the reduction in natural fire caused by fire control during the twentieth century (Curtis 1959; Lorimer 1985; Van Lear and Waldrop 1989). Periodic fires enable oaks to become dominant among competing species (Rouse 1986), and the oaks cannot replace themselves without disturbance (Kessler 1992).

Forest composition in the Southeast is also dominated by disturbance-initiated species and has been attributed to human activity over the last one or two thousand years (Buckner 1989). The exclusion of fire and major disturbance in many eastern forest systems is relatively recent and may now be producing completely new combinations of plant communities in certain regions. The abilities of animal species to adapt to these new plant communities are unknown. Clearly, the ways in which land is managed for particular uses will continue to influence the plants and animals that are present.

Serious environmental problems often come as surprises. Some of the more obvious symptoms of environmental degradation in the United States are now being addressed through public or private management actions. But what is the likely origin of the next environmental surprise? When examined in retrospect, some surprises have come from incremental shifts that crossed a threshold to catastrophic change. Such big effects from small causes (Ricker 1963) are known from many natural and social systems (Holling 1978). We suggest that it is the accumulation of small local changes that poses the greatest long-term challenge to encouraging sustainable land-use patterns in the United States.

Individual changes in land use may appear to have only local significance. In total, however, the large number of local changes is transforming the landscape of the United States. Gradual but widespread change leads to significant effects on vegetative cover, wildlife habitat, soils, and water quality. If the current path of land use in the United States continues, we can expect continued loss of wildlife and vegetation, erosion of soils, and nonpoint pollution of groundwater and surface water. If we are willing to manage landscapes for the larger good, these losses can be prevented, and the resources can be restored. Progress will require adaptive change and planning, monitoring, and research. Most importantly, it requires the political will to maintain a landscape that supports the natural resources we expect and need.

Present trends in land use suggest that collisions between the desires of the human population and natural ecological processes will continue. Ongoing urban development in our landscapes - including suburban sprawl and development of vacation homes - locks a pattern onto the landscape that is hard to reverse. Many farms of the previous century have been returning to more natural vegetation, but urban and suburban areas are more permanent. The larger lot sizes and the more sprawling nature of subdivisions that extend well into the countryside are an increasingly important cause of habitat loss and fragmentation. Housing located near natural areas may be quite desirable for the views and the privacy afforded; however, as a result of this practice, encounters between humans and wild animals or conflicts over the management of natural fires intensify. Developments are often designed to transport water off the land as quickly as possible, so developed watersheds become more prone to flooding. The fragmentation of large blocks of habitat can be minimized by the clustering of similar land uses, which has important implications for cities and suburbs. In general, it is more beneficial for habitat connectivity if human activities are concentrated than if they are dispersed widely throughout the landscape.

Land managers must enlarge their perspective to encompass the landscape. Clearly, plants and animals respond to regional patterns and to changes in habitat availability and connectivity. Natural disturbances, which create patchiness and structure biotic communities, must also be considered. Managers of natural areas that do not consider the surrounding seminatural environments lose crucial opportunities for maintaining more extensive habitats that harbor richer arrays of native plants and animals (Franklin 1993). When land managers focus too closely and don't "see the forest for the trees," the benefits of biological diversity may be lost. Management of freshwater habitat, for example, is likely to fail if the important riparian animals and plants are neglected or if land use in the floodplains is ignored. Land management should become a more cooperative venture that works within ecologically meaningful boundaries rather than political or jurisdictional boundaries.

Effective conservation of aquatic ecosystems and their biological diversity also requires a landscape perspective (National Research Council 1992). Unfortunately, most attempts to manage or restore aquatic systems address isolated components - for example, individual lakes, rivers, or wetlands. Agency responsibilities are oriented toward components rather than whole ecosystems (Leopold 1990), and the expertise needed for restoration is divided among various disciplines (National Research Council 1992). In practice, this division creates problems because uplands, wetlands, groundwater, rivers, lakes, and estuaries are interconnected by flows of water and nutrients and by migrations of organisms. These connections, though, must be considered. For example, wetlands are often essential to restoring lakes, but well-intentioned lake management (such as water level stabilization) could harm the wetlands. Pragmatic approaches to managing aquatic ecosystems require coordination at the watershed or landscape scale. At present, no organization or institution in the United States is responsible for the integrated view that makes watershed restoration practical (Cairns 1994). Addressing this challenge would benefit us all.

Species do not recognize political boundaries, suggesting the potential for considering some degree of integrated management of the natural and seminatural landscape. Of course, land-use regulation is a potentially explosive issue. A solution to balancing human needs and maintenance of biological diversity is not ecologically sustainable unless it is also politically sustainable. Forcing landowners to enhance biological diversity by regulation will probably produce a backlash that could lead to long-term loss of species. Research should proceed on the development of market-based incentives that encourage maintenance of species and habitats. Perhaps the greatest challenge for biological diversity protection is to preserve the structure and function of native ecosystems while protecting the rights and privileges of private property owners.

Many questions remain about the past and future effects of land use on animals and plants, and a few research needs stand out. Because the pattern of habitats across the landscape has such a strong influence on native animal and plant communities, improvements in methods for predicting future land-use patterns are sorely needed. Predicting land-use change requires linking knowledge and techniques from numerous disciplines, including economics, sociology, and ecology (Lee et al. 1992; Turner et al. 1996; Wear et al. 1996). Such interdisciplinary study is naturally complex, but there should be strong encouragement to develop the required integration. Better understanding is needed of what drives land-use change and how land-use patterns will respond to changing human population growth. Predictions of land-use change also should include not only the amount of change but also how the changes are arranged across the landscape. The effects of patch size and habitat fragmentation are sufficiently important that models should incorporate these factors (Turner et al. 1995). Models are a crucial component of research geared toward land use and biological diversity. Models permit us to project effects on plants and animals of the land-use experiments that society is presently conducting and to explore alternative future scenarios.

Greater explanation of what caused the pattern of plants and animals across today's landscapes is needed. The extent to which observed changes in present-day plant and animal communities are due to effects caused by changes in land-use practices decades ago or are directly related to recent practices is unclear. Some effects of past land use may persist for many decades, especially if long-lived organisms like trees are considered. In addition, the degree to which natural successional trajectories have been permanently altered by land use is not known. For example, the endpoints of the changing forest composition observed in the East and upper Midwest are not known. Many more empirical studies relating land-use change to biological diversity would be valuable, and more case studies would improve general understanding, as well as better explain local biotic patterns. These are important needs for relating land-use changes in the future to expected species' responses.

The most remarkable aspect of the landscape of the United States since European settlement is its continually changing face. Effects of these vast changes are long-lasting and crucial to our understanding of the present-day plants and animals that inhabit our landscapes (Foster 1992). In turn, conservation remains challenging in part because we seek to preserve areas that are changing (White and Bratton 1980). Ecological systems do not exhibit an undisturbed state that can be maintained indefinitely, but land use can alter both the rate and direction of natural trajectories. Because land-use patterns create the environment in which plants and animals must live, reproduce, and disperse, our understanding of the influence of land use on biological diversity is critical to the future of land management. The question is not whether we should or should not use land, but rather how we can best use the land. Open lands in the United States still abound in many areas. We are not predestined to continue past trends in land use. Rather, we have the opportunity to choose our destiny and plan for quality growth, balancing human needs while maintaining the integrity of our nation's primary resource - the land.

We especially thank J. Fraterrigo for her timely assistance with the library research required for this chapter. Comments from L. Iverson, J. Magnuson, and an anonymous reviewer improved the manuscript. This research was funded by the National Biological Service. The National Science Foundation also contributed to the ideas developed here through funding to the Coweeta and North Temperate Lakes Long-term Ecological Research Sites.