Courtesy F. Knopf, USGS

"The prairie, in all its expressions, is a massive, subtle place, with a long history of contradiction and misunderstanding. But it is worth the effort at comprehension. It is, after all, at the center of our national identity."

William Least Heat Moon (1991)

Williams and Diebel 1996">Estes et al. 1982ntribute immense value to watersheds and provide forage and habitat for large numbers of domestic and wild animals. Nevertheless, current levels of erosion in North America exceed the prairie soil's capacity to tolerate sediment and nutrient loss, thus threatening a resource essential to sustain future generations (Sampson 1981). Added to this threat is the potential for overgrazing by livestock and for other human activities to reduce the social and aesthetic values of grasslands and to restrict the commodities that grasslands can produce (National Research Council 1994), and the likelihood that severe degradation may be irreversible.

In North America, the prairie communities (about 1.5 million square kilometers) contain the majority of the continent's native grasslands (Fig. 1). Environmental features that describe the native North American grasslands embody similarity in vegetation, an abiotic environment to which the vegetation and structure respond, and the nature of the animal communities.

North American grasslands are similar in the general uniformity of their vegetation, dominance of grasses and grasslike plants, lack of shrubs, and absence of trees (Weaver 1968). The Great Plains grassland evolved in the rain shadow of the Rocky Mountains, where seasonal precipitation occurs mostly in spring and summer. From the Rocky Mountains east to the Mississippi River, the amount of precipitation increases and the frequency of droughts decreases (Simms 1988). Along a north-south gradient from central Texas to south-central Canada, the growing season becomes shorter, the average temperature decreases, and a greater proportion of annual precipitation occurs as snow. These broad-scale environmental gradients significantly influenced the evolutionary composition and distribution of prairie communities (Steinauer and Collins 1996; Weaver et al. 1996).

Many small and large grazing animals evolved on the North American prairie (Van Valkenburgh and Janis 1993), each with life-history and behavioral traits well adapted to the open character of prairie. For example, prairie dogs markedly affect the nutrient cycling, soil formation, and composition of grassland animal and plant communities (Miller et al. 1994). Moreover, it is important to realize that grasslands and their associated wildlife reflect events of the distant past (Knopf and Samson 1997). The incursion of animals--bison, elk, and others--into North America across land-bridges that once connected the Asian and North American continents is but one example of the role of past events. Thus, understanding how events of the distant past influenced both the isolation and interchange of plants and animals that interact with the more recent landscape is an essential step in interpreting any particular ecological community, including grasslands.

Understanding the biological resources of the Great Plains is difficult. The exact size of remaining grasslands in North America is unknown and difficult to estimate, and the community has undergone significant change since it was first described by early explorers and surveyors (Samson and Knopf 1994). Agriculture, urbanization, and mineral exploration have had both local and regional effects on biological resources. Invasions of nonindigenous plant species after fire suppression in the eastern, central, and southern prairies, as well as water developments in the western plains, have drastically altered grassland landscapes. Establishment of woodlots, shelterbelts, and tree-lined river and stream corridors within the prairie has contributed to a significant and ongoing loss of genetic diversity in North American grasslands (Knopf 1986).

This chapter highlights the status and trends of the main bodies of North American grasslands: the tall-grass prairie, the mixed-grass prairie, and the short-grass prairie. We feature the animals and plants dependent on native grassland and attempt to provide insight into the relationship between remaining native grassland and biological resources by reviewing available, current information and by describing threats.

Prairie Past and Present

In the past, grassland dominated central North America (Fig. 2) and, during the warm, dry interglacial times, reached--as the prairie peninsula--into parts of Wisconsin, Illinois, Indiana, and eastern Ohio (Bazzaz and Parrish 1982). The main bodies of native grassland, now vastly altered, are the tall-grass prairie extending from Canada (Manitoba) and Minnesota south to Texas; the mixed-grass prairie from Canada, Montana, and North Dakota south to Texas; and the short-grass prairie extending from eastern Wyoming south to western Texas and eastern New Mexico. In the north, the natural grasslands are bordered on the west by coniferous forests of the Rocky Mountains and on the east by oak savannah (Anderson 1983) and aspen parkland in Manitoba and northwest Minnesota, with the transition from prairie to forest often abrupt (Great Plains Flora Association 1986). Across the Great Plains, coniferous and deciduous forest types meet only in the valley of the Niobrara River in north-central Nebraska, and isolated stands of both forest types occur in the Black Hills of South Dakota.

Tall-grass prairie (Fig. 3) is the wettest of the grassland provinces and is predominantly composed of sod-forming bunch grasses. Like other grasslands, the tall-grass prairie has species originally from different geographical sources (Simms 1988). Grassland groupings of the tall-grass prairie are the bluestem prairie from southern Manitoba through eastern North Dakota and western Minnesota south to eastern Oklahoma, and the wheatgrass, bluestem, and needlegrass area from south-central Canada through east-central North Dakota and South Dakota to southern Nebraska.

Three additional areas are associated with tall-grass prairie: the Crosstimbers, a band of grassland and oak savanna at the southern edge of the bluestem prairie in Kansas to the Trinity River in Texas (Küchler 1964), the Blackland Prairie south of the Crosstimbers (Gould 1962), and the rice prairies. The rice prairies are former coastal prairies that have been converted to rice production (Hobaugh et al. 1989). The original vegetation in rice prairies was mainly tall grass and extended across 9,000 square kilometers, largely along the Texas coast and inland as much as 125 kilometers, and into Louisiana. Little coastal prairie remains; Attwater's Prairie Chicken National Wildlife Refuge in Texas is the single major remnant.

Since 1830 declines in the area of tall-grass prairie within specific states and provinces are estimated to be 82.6 to 99.9% (Table 1) and exceed those reported for any other major ecological community in North America (Samson and Knopf 1994). Iowa, for example, has barely 12,140 hectares remaining of its original 12 million hectares of tall-grass prairie. Less than 1% of the presettlement tall-grass prairie remains in Manitoba, Illinois, Indiana, and North Dakota. Minnesota and Missouri, states active in prairie conservation, work with less than 9% of the presettlement tall-grass prairie. Tall-grass prairie remains important to ranching in the Osage and Flint Hills of Kansas and in tracts in South Dakota, Oklahoma, and Texas.

One can envision the short-grass and tall-grass prairies intergrading just east of an irregular line that runs from northern Texas through Oklahoma, Kansas, and Nebraska, northwestward into west-central North Dakota and South Dakota (Figs. 1 and 2). The perimeter is not well defined because of the array of short-stature, intermediate, and tall-grass species that make up an ecotone between the short-grass and tall-grass prairies (Bragg and Steuter 1996). In general, the mixed-grass prairie is characterized by the warm-season grasses of the short-grass prairie to the west and the cool- and warm-season grasses, which grow much taller, to the east (Fig. 4). Because of this ecotonal mixing, the number of plant species found in mixed-grass prairies exceeds that in other prairie types. Estimated declines in area of native mixed-grass prairie, although less than those of the tall-grass, range from 30.5% in Texas to over 99.9% in Manitoba (Table 1).

The short-grass prairie extends east from the Rocky Mountains and south from Montana through the Nebraska panhandle and southeastern Wyoming into the high plains of Oklahoma, New Mexico, and Texas (Figs. 1 and 2.). The short-grass prairie landscape (Fig. 5) was one of relatively treeless stream bottoms and uplands dominated by blue grama and buffalo grass, two warm-season grasses that flourish under intensive grazing (Weaver et al. 1996). Buffalo grass reproduces both sexually and by tillering sprouts from the base of grass clumps. Unlike the more eastern species, short-grass prairie species remain digestible and retain their protein content when dormant.

Declines in short-grass prairie have generally been much less than those of tall-grass and mixed-grass prairies (Table 1). However, perhaps in no other system than short-grass prairie are historical and evolutionary impacts of grazing so apparent (Knopf 1996). Clearly, birds endemic to the short-grass prairie express life-history characteristics and habitat use in response to grazing (Fig. 6). The mountain plover responds to highly disturbed sites, the chestnut-collared longspur to moderately grazed areas, and the Baird's sparrow to sites with taller grasses. In the mid-1800's the numbers of individuals of native mammal species--bison, prairie dogs, pronghorn, elk, grizzly bears, and gray wolves--rivaled or exceeded those now in the African Serengeti (Howe 1994). Major antigrazing structures evolved in plants: thorns and spikes; thick or hard tissues difficult to bite, chew, or digest; and secondary compounds difficult to digest. These structures have arisen through the long coevolutionary association between plants and animals with grazing on grasslands.

At present, extensive areas of short-grass prairie are dominated by invasive perennial and annual species, whose presence is attributed to overgrazing by domestic livestock and dryland farming (Weaver et al. 1996). To the south, specifically the Texas high plains, much of the short-grass prairie is now farmland or shrubland invaded by prickly pear cacti and oaks. Only the short-grass prairie and, to a lesser extent, mixed-grass prairie remain in public ownership. These areas are largely on the national grasslands managed by the U.S. Forest Service.

The northern prairie contains numerous wetlands (Fig. 7), including the glaciated prairie pothole region (Fig. 8), the Nebraska Sandhills, and the Rainwater Basin (Kresl et al. 1996; Mack 1996). The 770,000- square-kilometer prairie pothole region extends from Alberta, Saskatchewan, and Manitoba across northeastern Montana, then southeast through North Dakota and eastern South Dakota into western Minnesota and northwestern and central Iowa. This landscape is pockmarked with numerous small, shallow depressions that capture snowmelt and rainwater or are within reach of subsurface waters. Estimated losses of prairie pothole wetland range from 35% in South Dakota to 99% in Iowa with loss rates upwards of 1,300 hectares per year (Tiner 1984).

The Nebraska Sandhills is the largest dune area (5,260 square kilometers) in North America (Mack 1996). About 526,100 hectares of wetlands are scattered throughout the area. The rapid movement of groundwater creates a continuum among lakes, wetlands, and streams, thus an alteration in one area may easily affect vegetation and wetlands over a larger landscape. Wetlands in the sandhills range from shallow, extremely alkaline basins to deeper, freshwater lakes to spring-fed streams. They are economically valuable, particularly as a source of irrigation water.

Another wetland area dependent on capturing water runoff is the Rainwater Basin of south-central Nebraska (6,720 square kilometers). Throughout the basin, rainwater is caught by scattered wind-excavated depressions underlain by an impermeable clay pan. Since the late 1800's efforts have been under way to drain Rainwater Basin; today, fewer than 400 depressions remain of an estimated 4,000, and they account for 22% of the former area. Other large, alkaline wetlands in Kansas include the Jamestown marsh, Talmo marsh, Lincoln salt marsh, Cheyenne Bottoms, Quivira National Wildlife Refuge (including Big and Little Salt marshes), and Slate Creek salt marsh. A similar alkaline lake is the Great Salt Plains Reservoir in Alfalfa County, Oklahoma.

Effects of collective water loss on the northern prairie range from significant declines in waterfowl breeding populations (Bethke and Nudds 1995) to elimination of the flood storage value of natural wetlands. About half of the continental waterfowl production comes from the prairie pothole region. Nebraska's Rainwater Basin is the major spring staging area for the buff-breasted sandpiper and the greater white-fronted goose, and it provides migratory habitat for endangered species such as the whooping crane and interior least tern. In addition, about 45% of North America's shorebird population east of the Rocky Mountains may stop at Cheyenne Bottoms during spring migration, including 90% of the North American populations of the white-rumped sandpiper, stilt sandpiper, Baird's sandpiper, long-billed dowitcher, and Wilson's phalarope, and over half of all pectoral sandpipers, marbled godwits, and Hudsonian godwits.

Interior wetland from the edge of the prairie pothole region across the central Great Plains is associated with major river systems. The area has few natural lakes, the largest of which is Inman Lake (78 hectares) in central Kansas (Carlander et al. 1986). Climate and past events account for the interior wetland's habitat characteristics (Cross and Moss 1987). On the central Great Plains, such areas have been transformed from wide, unvegetated channels (Fig. 9) to the current extensive cottonwood- willow woodlands lining narrow channels (Johnson 1994). This transformation is a result of human alteration of natural flow regimes, cessation of prairie fires, and elimination of the bison (Currier 1982). Overall, the presettlement near-river mosaic of meadow, marsh, and drier upland grassland is now an open- to closed-canopy woodland. Major changes in the habitat mosaic are expected in the future because secondary succession of woody vegetation will lead to a climax forest of regenerating stands of the nonindigenous Russian-olive along the Platte River and elsewhere across the entire central and western United States (Olson and Knopf 1986).

To the south are the Playa lakes. Upwards of 26,000 of these shallow, small wetlands dot the short-grass prairie in Texas and adjoining states, reaching northward to Colorado (Bolen et al. 1989). Because of evaporation, such lakes often are rimmed in a salty crust. Playa lakes have abundant populations of small, aquatic invertebrates (Carlander et al. 1986). At times, these lakes serve as vital waterfowl nesting and resting areas (U.S. Fish and Wildlife Service 1986) and as winter habitat for most North American sandhill cranes.

The High Plains aquifer (formerly called the Ogallala aquifer) consists of one or more geological units connected belowground under the central Great Plains. The aquifer is essential to agricultural, urban, and environmental resources, containing about 20% of the irrigated farmland in the High Plains and about 30% of the water used for irrigation (Huntzinger 1996). Precipitation is the principal source of natural groundwater recharge, but recharge can also result from seepage loss from streams and lakes. Natural discharge occurs as evaporation from plants and soils where the water table is near the surface or as seepage to springs. Over the long term, natural recharge should compensate discharge.

The development of the High Plains aquifer for irrigation (1940-1980) is evident in an average area-weighted, water-level decline of 3 meters (0.07 meters per year; Dugan et al. 1994). Declines vary with locale, exceeding 30 meters in some parts of the central and southern High Plains; 6 meters in southwestern Kansas, east-central New Mexico, and the Oklahoma and Texas panhandles; and 3 to 6 meters in northeastern Colorado, northwestern Kansas, and southwestern Nebraska. Since 1980, water levels in such areas have continued to decline but at a slower rate, the result of greater than normal precipitation (10 to 15 centimeters annually), water conservation practices (particularly minimum tillage), and reduction in irrigated land area (about 540,000 hectares, 1979 to 1991). The extreme southern plains of Texas and eastern and east-central Nebraska have experienced water-level rises of 1 to 3 meters (Dugan et al. 1994).

Agricultural use of nitrogen fertilizers is the largest source of nitrates in near-surface aquifers in the midcontinent (Koplin et al. 1994). Over 100,000 metric tons of pesticides (herbicides, insecticides, and fungicides) were applied in the midcontinent in 1991, often to control nonindigenous plants and animals. In spring and summer 1991, concentrations of several herbicides exceeded U.S. Environmental Protection Agency standards in about half of the streams sampled in the upper Missouri River basin (Huntzinger 1996). Effects of these pollutants on the quality of human life and on the integrity of the ecological community are largely unknown. The U.S. Environmental Protection Agency has initiated an effort to develop stressor information to help recognize areas where urban development, agricultural nonpoint pollution (pesticides, toxic chemicals, nutrient pollution), and agricultural development may exacerbate ecological decline.

The rivers of the Great Plains flow from west to east; extreme turbidity is the key characteristic of the larger rivers. Whereas most water in these rivers originates from western mountains, sediments originate from thunderstorm runoff on the Great Plains. In small channels, fine particles held in suspension produced quicksands, which inhibited crossings by early travelers and caused extreme turbidity during low flows. In summer, open-river water temperatures often exceed 30°C, the salinity level is high because of salt- and gypsum-laden groundwaters, rates of evaporation are high, and the flow velocity is moderate.

The larger Great Plains rivers have been subjected to dewatering for irrigation, other consumptive uses, and reservoir construction (Cross and Moss 1987). In virtually all these river systems, such dewatering has altered the timing and extent of flows, downstream temperatures, levels of dissolved nutrients, sediment transport and deposition, and the structure of plant and animal communities. Few major interior rivers still exhibit the conditions evident before agricultural development and water management had occurred.

On the Missouri River, once a free-flowing river of more than 3,700 kilometers (from Three Forks, Montana, to St. Louis, Missouri), reservoir construction in Montana, North Dakota, and South Dakota has virtually stopped sediment transport below major reservoirs, as it also has on the Platte and Kansas rivers (Huntzinger 1996). Sediment deposition is part of reservoir design but remains a maintenance concern. Because of the value of surface waters, there is increasing interstate interest in surface waters such as those of the Missouri River. Such surface waters are important to wildlife, fish, and recreation, as well as to navigation and water-supply interests that rely on reservoir resources.

Ecologically, the effect of water management reaches far to the south because sediment deposits are needed to sustain the Mississippi Delta. Unchannelized reaches of the Missouri River are near Bismarck, North Dakota and in southeastern South Dakota and northeastern Nebraska. Such reaches host a number of pearlymussel, fish, and bird species that are federally listed or are candidates for listing under the Endangered Species Act of 1973.

The basic features of most Great Plains streams, such as flow and substrate, are unknown (Matthews 1988). In general, streams in the south are characterized by irregular flow, small particle size in substrates, and a distinct wet-dry cycle. Drought may be a more extreme disturbance than either the wet-dry cycle or floods. Streams in the northern prairies are more consistent in flow and tend to have cobble substrates, and winter precipitation (snow) is released with spring thaw. Great Plains streams fall into three categories: the shallow stream with shifting sand beds; clear brooks, ponds, and marshes supported by seeps and springs; and residual pools of intermittent streams (Cross and Moss 1987). All three stream types are affected by dewatering of channels and stabilized flows.

Grassland soils are fundamentally different from those found under a forest canopy (Simms 1988). Many factors--parent material, climate, time, and human intervention--influence grassland soil type and condition (Peterson and Cole 1996). For example, soil organic carbon, a significant influence on soil productivity, is greatest in the Northeast (cooler and wet) and lower in the Southwest (warmer and dry). Where evaporation is low, water is more likely to remain in the soil, increasing the rate of mineral weathering and allowing large amounts of nitrogen, phosphorus, and sulfur to accumulate in conjunction with carbon.

Managed agriculture began on the easternmost grasslands in the 1850's (Peterson and Cole 1996). Surface cover is reduced by agriculture, and soil structure is destabilized by reducing aggregate size--the result of mixing and grinding action by farm implements. Organic carbon loss is accelerated by agriculture, and cultivated crops (particularly in dryland areas) return little carbon to the soil. Few agricultural practices of the early settlers captured or retained moisture. The black-dust storms of the 1930's resulted from exposing vast areas of cultivated prairie soil to wind action and drought (Sampson 1981).

Several straightforward and well-documented relationships exist between plant productivity and soil organic material after native sod is cultivated. For example, soil productivity (indexed by corn grain yield) declined 71% and soil nitrogen 49% during a 28-year interval after cultivation began (Williams and Wolman 1986). Retention of organic matter--and thus the level of productivity--in grassland soils is only possible if the correct proportions of carbon, nitrogen, and phosphorus are present (Peterson and Cole 1996). Nitrogen fertilizers are used extensively to restore soil nitrogen levels, and more than 6.4 million metric tons of nitrogen fertilizers were applied to cropland in the Mississippi River basin in 1991 (Goolsby et al. 1993). In addition, removal of phosphorus in harvested plants and loss of organic phosphorus due to cultivation require fertilizer supplements to maintain productivity. Elevated concentrations of phosphorus may affect aquatic plant growth and reduce oxygen content in streams.

Soil formation is a slow, continuous process. About 2.5 centimeters of new topsoil is formed every 100 to 1,000 years, depending on climate, vegetation and other living organisms, topography, and the nature of the soil's parent material (Sampson 1981). Some soils, particularly where moisture is a limiting factor and growing seasons are short, may take 10,000 years to produce 2.5 centimeters of soil. On average, annual loss of topsoil in the United States is nearly three times greater than that being formed. Exact measurements of soil losses are difficult to obtain because soil eroded in one area is eventually deposited at another site (Peterson and Cole 1996). Smaller and lighter nutrient-rich organic soils are the most transportable, creating additional threat to the future productivity of prairie soils.

Climate and fire (Fig. 10) are thought to be most important to the spread and maintenance of grasslands (Anderson 1990). The air mass originating in the Gulf of Mexico spreads high humidity and precipitation as it moves north. As it moves from west to east, the Pacific air mass passes over several mountain ranges, giving up most of its moisture and, as it meets the gulf air mass, creating a climate gradient across the Great Plains. The north-south gradient in temperature and moisture, largely influenced by snow cover to the north, is an effect of the polar air mass. Long-term response of vegetation to climate, particularly water availability, is illustrated by regional differences in species composition and height of native grasses: arid western short-grass prairie, central mixed-grass prairie, and eastern tall-grass prairie. As documented in past droughts, grassland distribution is controlled by extremes of climate variability. These forces are evident in the changing eastern edge of the prairie peninsula and, to a great extent, the annual productivity of grasslands.

Grassland plants have growing points protected from fire beneath the soil surface.

Frequent fire is essential to maintaining native species diversity, and it affects other components, including nutrient cycling and productivity (Collins and Wallace 1990). On tall-grass prairie, the relationship between total plant species richness and the number of times a site is burned is important and positive (Collins 1991), at least on a small scale. Small animals create gaps and edges that influence tall-grass plant community structure and composition (Reichman et al. 1993). Recently burned tall-grass prairie has also provided stopover habitat for many long-distance migrant birds such as the lesser golden-plover and the now-endangered and possibly extinct Eskimo curlew. In the past, grazing was localized in these burned areas because of their greater productivity and the nutritive value of their forage (Risser 1990). Thus, the movement and impact of grazing animals on tall-grass prairie grasslands were bound to the spatial distribution of burned patches. For mixed-grass prairie, discussions of community composition and individual species should be set in a similar context of species patterns of both past grazing animals (ranging from bison to ants) and current grazing animals (domestic livestock) in relation to disturbance events (Umbanhowar 1992).

Grazing has direct and indirect effects at landscape and regional scales, which, in turn, interact with other small-scale and large-scale factors to heighten temporal and spatial diversity in grasslands (Gibson and Hulbert 1987; Risser 1990). A recent comparison of grazing over a global range of environments, however, suggests grazing is a factor in the conversion of grasslands to less desirable shrublands (Milchunas and Lauenroth 1993). Moreover, primary production on grasslands, largely the production of plant material, does not necessarily change when plant species composition changes. Current species-based management criteria by land management agencies, therefore, may lead to erroneous conclusions about the ability and future of grasslands to sustain productivity. Adequate assessment of the effects of grazing on grasslands, as with the effects of climate and fire, must be multiscaled and match management inferences and applications (Steinauer and Collins 1996).

Interactions among other factors, aside from climate, grazing, and fire, also influence grasslands (Burke et al. 1991). In the east, nitrogen normally restricts the annual production and composition of grasslands. In the semiarid west, the availability of nitrogen and water is important to composition and production. Long-term vegetative production on short-grass prairie is closely tied to precipitation (Lauenroth and Sala 1992). The most productive years are those when small precipitation events first stimulate nutrient availability, followed by large precipitation events that stimulate plant growth. Semiarid areas are thought of as especially variable in environmental conditions, particularly in precipitation, and the short-grass prairie is no exception. Effective grassland management requires understanding the effects of both the spatial and temporal patterns of precipitation on short-grass prairie.

In prairie wetlands, disruption of natural processes such as fire has led to domination by robust, emergent plants, particularly in the prairie pothole region. Cattail, once rare on the Great Plains, has spread across thousands of prairie wetlands, as has purple loosestrife, a species native to Europe which is now threatening waterways across the United States (U.S. Congress, Office of Technology Assessment 1993; Malecki and Blossey 1994). In the past, climate, fire, and grazing controlled the diversity and abundance of vegetation in northern prairie wetlands. As environmental conditions changed, some plant populations have declined and others have increased. Belowground seed reserves favor those species with seeds that germinate under a wide range of conditions, such as cattail, purple loosestrife, and other nonindigenous species.

More is known about the effects of grazing than fire. Nodal rooting, or underground branching, and unpalatability are evident evolutionary responses of wetland plants to grazing. Under certain conditions grazing can increase species diversity and the development of intricate patterns and sharp boundaries among prairie wetland plant communities (Bakker and Ruyter 1981).

The Nature Conservancy, in a preliminary survey, has identified rare plant assemblages across the United States (Grossman et al. 1994). Of the 633 assemblages in the Great Plains, 107 (17%) are considered rare (Chaplin et al. 1996).

The 16 rare Great Plains forest assemblages are largely cottonwood and oak floodplain forests on the eastern and western edges of the plains (Grossman et al. 1994). The 20 rare canyon and mountain plant assemblages tend to be open pine, fir, and oak. The eight rare sparse woodland forests are primarily oak savannas on the eastern plains. The 19 rare shrubland assemblages include many sagebrush, hawthorn, and willow species.

Among 45 rare grassland assemblages on the Great Plains, 18 are found in tall-grass prairie, 13 in mixed-grass prairie, 7 in short-grass prairie, and 7 primarily in wetlands. Big bluestem is dominant in 9 of 18 rare tall-grass prairie communities, little bluestem in 3, and drop-seed species in 2. Similarly, little bluestem is common to 6 of 13 rare mixed-grass prairie communities, and sedges are important in 3. Buffalo grass, in part, distinguishes 5 of 8 rare short-grass prairie communities, and sedges are important to 2. The 7 remaining rare communities are dominated by forbs and embrace wetland plants.

A wide diversity of terrestrial insects exists on grasslands. For example, in two years of sampling on a 1,400-hectare area of tall-grass prairie in northeastern Oklahoma, 16 orders, 131 families, and more than 3,000 insect species were noted (Risser et al. 1981). More than 1,600 insect species are known from a short-grass prairie in Colorado (Kumar et al. 1976), and this list is incomplete. Inventories are rarely representative; some taxa are present in hot, dry years, others in wet years, and no single sampling method is adequate.

Other terrestrial invertebrates are also abundant. Smolik (1974) found 2 to 6 million soil nematodes per square meter to the depth of 60 centimeters in South Dakota mixed-grass prairie soil. Terrestrial invertebrates are important to the prairie community: they feed on plant tissue, pollen, nectar, and seeds; regulate numbers of other insects and plants; and recycle energy and nutrients (Risser et al. 1981). Underground, earthworms accelerate the decomposition and mineralization of soil organic matter and affect soil structure through burrowing and casting. The soil formation activities of native and nonindigenous earthworms vary considerably; the latter have a negative effect on soil turnover, at least in tall-grass prairie soils (James 1991). Nevertheless, in short-grass prairie soils, 90% of invertebrate energy cycling occurs belowground, less in tall-grass and mixed-grass prairies.

Preliminary lists of Lepidoptera (butterflies and moths) are available for Montana, North Dakota, Wyoming, Colorado, and New Mexico and are in progress in Texas (Powell 1995). Species numbers in a few selected Lepidoptera families vary from 181 in North Dakota to 520 in Texas, and numbers in Nebraska (254) and Oklahoma (228) rank high (Opler 1995). Several prairie butterflies--the Dakota skipper (Fig. 11), regal fritillary, tawny crescent, and maculated manfreda skipper--and two moths--the rattlesnake-master borer moth and phlox moth--are federal species of concern. An additional six insects--the persius duskywing, poweshiek skipperling, ottoe skipper, byssus skipper, silver-bordered fritillary, and Ozark emerald--are considered species of concern by The Nature Conservancy, and another six are endemic to the prairie (Royer 1992).

Adequate inventory and distribution information is unavailable for predicting status and trends for most invertebrates. Ranges of a number of grasshopper species seem focused specifically on short-grass, a combination of short-grass and mixed-grass, and tall-grass prairies (Otte 1981). Ranges of many other grasshopper species center on grasslands but extend into adjacent forested and scrub habitats. The prairie mole cricket, a strikingly large insect (Fig. 12), and the superb spharagemon grasshopper are federal species of concern. Additional species of concern are the Ozark snaketail dragonfly, a true bug, a fly, six amphipods, nine cave spiders, two cave beetles, a cave amphipod, a cave shrimp, and a number of beetles, including the widely distributed sixbanded longhorn beetle.

Leafhoppers are among the most diverse and well-studied terrestrial insects on the grasslands (Whitcomb et al. 1994). The ranges of some species span considerable distances across the prairie, whereas other species are more restricted in distribution. Most are highly specialized, often endemic, and good indicators of grassland condition. On the short-grass prairie, 20 leafhopper species may colonize a single host, such as blue grama, which indicates a long record of coevolution. There is, however, a rather spectacular partitioning of resources: no more than 10 leafhopper species occur on the blue grama host at a given site. The "invisible wall" partitioning resources is climate, with each taxon showing singular limits to humidity, cold, and heat. Understanding any single assemblage of leafhoppers (and perhaps most other insects) requires knowledge of the communities in which they and their relatives reside and of the past structure and conditions of these communities.

Around 1889 abundant mussel populations of the Great Plains were recognized as an economic resource, particularly as the materials for button production. One example of the abundancy of this resource was a single mussel bed that covered an area of 2.4 kilometers by 288 meters in the Mississippi River near New Boston, Illinois (Carlander et al. 1986). The mussel bed was depleted by 1898, and large-scale propagation to restore the resource failed. No federal regulations restrict the harvest of mussels unless they are federally listed as endangered or threatened. Many states (Illinois, Iowa, Kansas, Minnesota, Missouri, Montana, Nebraska, North Dakota, Oklahoma, South Dakota, and Texas), however, have instituted harvest regulations.

Along with siltation and contamination, dams, with their altered flow regimes and accompanying reservoirs, are believed to have caused declines of mussels and other aquatic organisms. Four pearlymussels--the elktoe, spectacle-case, snuffbox, and scaleshell--are federal species of concern. The American Fisheries Society has identified 213 of 297 (71.7%) mussels found in the United States and Canada as threatened, endangered, or of special concern (Williams et al. 1993). Many of these species are endemic; for example, all three endangered and six threatened species, and about one-third (two of seven) of the mussels of special concern recently added by the American Fisheries Society list for Texas are species with restricted ranges. Several nonindigenous species, particularly the Asian clam and the zebra mussel, pose potential threats to the native mussel fauna on the Great Plains.

Across the grasslands, a number of snails, including both land and aquatic species, are federal species of concern: 10 are found in Texas, at least 8 in the upper Midwest (Illinois, Iowa, Minnesota), 5 in the central plains (Missouri and Oklahoma), and 3 in the West (Wyoming and the Dakotas). One snail, the Iowa Pleistocene snail, is listed as endangered. The same factors negatively affecting other species--habitat loss, drainage, water pollution, stream desiccation accompanying the lowering of water tables for agriculture and municipalities, and competition with and predation by introduced species--are thought to threaten freshwater native snails, crayfishes, and other aquatic invertebrates.

In general, drainages east of the Rocky Mountains have a richer fish fauna than those to the west (Brooks and McLennan 1993). The western Mississippi basin occupies much of the grasslands, including the basins of the Missouri, Arkansas, and Red rivers. Geological and presettlement history and current information depict changing fish distributions and abundances across this vast region (Cross et al. 1986). Geological information provides insights into how glacial advances fragmented fish populations into smaller, subregional distributions and transported species to distant and distinct drainages. Government surveys seeking routes for human immigrants to the Southwest and Pacific coast provide important historical information. Recent concern for prairie (and other) fishes has encouraged the continuation of many ongoing surveys.

The western Mississippi basin contains at least 266 species of fishes from the United States, which is slightly more than one-third of the fish species in the United States and Canada (Cross et al. 1986). Thirty-one species were introduced to North America or now exist beyond their past ranges. Two species are diadromous (and thus reside at some time in their life cycles in marine waters), and 34 are endemic to one or more drainages in the vast basin.

Cross et al. (1986) described two physiographic regions that encompass the habitat of most fishes on the grasslands: the Great Plains, from eastern Montana south to western Texas, and the Central Lowlands, from North Dakota south to eastern Texas. Thirteen fishes are endemic to the two physiographic regions: 2 to the Central Lowlands, and 11 to the Great Plains. The Great Plains province has 77 fish species (86 including introduced species), that, with the exception of 5 species, are a subset of the Central Lowlands fish assemblage. The Central Lowlands has 139 native species with over 24 species introduced into the region, most intentionally as sport fish or forage for sport fish. Eleven grassland fishes are large-river species that have centers of origin in one or two adjacent physiographic regions but are shared among the Central Lowlands and Great Plains physiographic regions. Overall, the species inhabiting small streams and rivers outnumber those in large rivers about ten to one.

Decades of intensive agricultural development and modified flow regimes are held responsible for declines in the fishes endemic to the small streams and turbid rivers of the Great Plains (Cross and Moss 1987). Nevertheless, the first declines noted in regional endemic fishes were in those species in small, clear, spring-fed streams, particularly streams that were home to the Topeka shiner and Arkansas darter. Although agriculture and altered flow regimes may explain declines, patterns in decline differ among species, with colonization of suitable habitat of restricted stream and river reaches exceptionally rare. Several fish species of shifting-sand bottom streams in the Great Plains are considered federal species of concern, including the sicklefin chub, the sturgeon chub, the Arkansas River speckled chub (a subspecies of the speckled chub), and the Arkansas River shiner. The Topeka shiner, the plains topminnow, and the Arkansas darter, which occur in clear streams fed by springs and seeps, and the lake sturgeon are also candidates for species of concern (Echelle et al. 1995).

Overall, 9 of 10 species and subspecies of broad, shallow, and sandy-bottom streams seem to be in serious decline. Two additional species, the plains minnow and flathead chub, seem to be drastically declining (V. M. Tabor, U.S. Fish and Wildlife Service, Manhattan, Kansas, unpublished manuscript) and are candidates for federal listing. Marked increases are known in some fishes in the natural prairie community. Clear-water fishes, particularly sunfish, perch, and introduced species such as carps, replace species native to turbid conditions and tend to be increasing. Within the Missouri River drainage, the upper Missouri and Yellowstone rivers provide the best remaining habitat for the large-river natives. About 20% of the native large-river fish species are declining.

Most grassland reptiles and amphibians are widely distributed. Half of the species (6 of 12) found in Alberta also occur several hundred kilometers to the south in northern Mexico (N. J. Scott, Jr., U.S. Geological Survey, San Simeon, California, unpublished manuscript). More than 40 reptiles and amphibians are characteristic of prairie habitat (Table 2). The number of local species is influenced by the presence of water (which amphibians need to complete their life cycle), complex habitats, and sandy or loose soils needed for concealment by some species.

Loss of small water areas, nonindigenous terrestrial and aquatic predators, grazing, exotic plantings, and prairie dog control are believed to contribute to reptile and amphibian declines. Most reptiles and amphibians rely on temporary ponds rather than streams or rivers. Permanent water provides habitat for the especially predatory bullfrog, catfish, and sunfish. Woody vegetation near permanent water favors mammalian predators such as the Virginia opossum, raccoon, and skunk (Schwalbe and Rosen 1989). Moderate grazing increases habitat structure and patchiness important to reptile and amphibian abundance, but overgrazing reduces needed habitats, as does planting of nonindigenous species such as buffelgrass (Scott 1997). Prairie dog burrows provide winter retreats and summer nesting sites for reptiles and amphibians, thus their destruction may cause local reptile and amphibian declines.

Most grassland reptiles and amphibians seem widespread and secure. Several, however, particularly those with very restricted ranges, are thought to be declining. Habitat for the reticulate collared lizard and spot-tailed earless lizard, species restricted to south Texas, is threatened by exotic buffelgrass (Scott 1997). Agriculture in the Lower Rio Grande valley in Texas has reduced the number of temporary ponds, which are needed by the black-spotted newt. Pesticides may negatively affect the Wyoming toad, an endangered species, but conclusive evidence is lacking. Across its range, the yellow mud turtle, a federal species of concern, is restricted to a few widely distributed ponds (Dodd 1983).

Of the 435 bird species that breed in the United States, 330 breed on the Great Plains (Knopf and Samson 1995). Nevertheless, few North American bird species are believed to have evolved within the Great Plains. Mengel (1970) suggested that only 12 bird species are endemic to the grasslands. An additional 25 species are believed to have evolved on the grassland, though they range widely into adjoining vegetation provinces. Five of these 25 species are specifically associates of sagebrush landscapes of the Great Basin (Knopf and Samson 1995).

As a group, the endemic grassland bird species have shown more consistent, widespread, and steeper declines (Table 3) than any other guild of North American bird species (Knopf 1992, 1996). Individually, populations of the mountain plover (Figs. 13 and 14a), Cassin's sparrow (Fig. 14b), and clay-colored sparrow (Fig. 14c) are declining throughout their breeding ranges. Breeding habitats are disappearing locally for the Franklin's gull (Fig. 14d), the dickcissel (Fig. 14e), the Henslow's sparrow (Fig. 14f), the grasshopper sparrow (Fig. 14g), and the western meadowlark (Fig. 14h). Breeding ranges are shifting for the ferruginous hawk (Fig. 14i), the Mississippi kite (Fig. 14j), the upland sandpiper (Fig. 14k), the horned lark, the vesper sparrow, the savannah sparrow, and the Henslow's sparrow. Populations of the wetland-associated marbled godwit and Wilson's phalarope seem stable, and populations of the upland sandpiper and McCown's longspur (Fig. 14l) have increased markedly.

Most of the endemic bird species of the short-grass and mixed-grass prairies are associated with large grazing animals (Fig. 15); others, such as the ferruginous hawk, prairie falcon, and burrowing owl, are either somewhat or strongly associated with prairie dog colonies.

Locally, drought tolerance seems to be the principal ecological process influencing the composition of grassland bird communities (Wiens 1974), with grazing (Hobbs and Huenneke 1992) and wildfire (Zimmerman 1992) playing secondary roles. Unlike most forest birds that winter in the Neotropics, virtually all endemic grassland birds winter on the continent; thus influences on their status and trends occur within North America.

The prairie pothole region is a key breeding area for species such as the mallard, blue-winged teal, and northern pintail. Researchers believe that preserving native grassland and wetlands is essential for slowing declines in duck numbers, including the mallard, American widgeon, and northern pintail (Canadian Wildlife Service and U.S. Fish and Wildlife Service 1994). Predation on eggs and hatchlings by red foxes, striped skunks, raccoons, and other species substantially reduces the abundances of ducks (Ball et al. 1994).

The grasslands have a fertile history of plant- and seed-eating mammals (Hall and Kelson 1959). A rich array of large herbivores, including camels, rhinoceroses, mammoths, mastodons, and bison, evolved on the North American grasslands only to disappear during the last Ice Age (Van Valkenburgh and Janis 1993). The surviving large grassland herbivore, the plains bison, evokes a mystique not shared by any other North American mammal and which is largely derived from Native American and frontier heritages (Meagher 1978). In the past, bison (Fig. 15) numbered from about 60 to 70 million and roamed in large herds that, in the 1860's, often required horseback riders several days to successfully penetrate and cross.

As many as 5 billion prairie dogs may have been in North America before European settlement. An estimated 98% decline in prairie dog numbers has occurred since European settlement (Summers and Linder 1978). The black-tailed prairie dog may occupy less than 0.5% of its original range, the short-grass and mixed-grass prairies. As a result, a variety of species closely associated with the prairie dog are either federally listed as endangered or are being considered for listing as threatened or endangered. These include the black-footed ferret, the swift fox, and the mountain plover.

The ranges of more than 100 native mammals extend into the prairie; nearly half of these occur in the forest-grassland ecotone and others in diverse habitat types (Risser et al. 1981). Nevertheless, surveys of eastern and western species whose ranges stop short of the Great Plains support the effectiveness of the grassland as an evolutionary barrier to dispersal (Hall and Kelson 1959). Estimates of Great Plains-restricted mammals range from as few as 10 (Risser et al. 1981) to as many as 18 (Jones et al. 1985). These species include one lagomorph (the white-tailed jack rabbit), eight rodents (thirteen-lined ground squirrel, Franklin's ground squirrel, black-tailed prairie dog, plains pocket gopher, olive-backed pocket mouse, plains pocket mouse, plains harvest mouse, and prairie vole), and two carnivores (swift fox and black-footed ferret).

One recent extinction, the Audubon bighorn sheep, was a subspecies found along the upper Missouri River, including North Dakota, South Dakota, Wyoming, and Nebraska. Populations of the caribou, a species once common across northern North Dakota but which is now extirpated, are still common in Canada. Similarly, the gray wolf and elk, once common on the grasslands, and the less common mountain lion and wolverine are found elsewhere.

Prairie Integrity and Legacies

Integrity here means maintaining species and ecological processes characteristic to a particular landscape (Samson and Knopf 1993); this is an emerging goal in resource conservation (Angermeier and Karr 1994). The human-caused breakdown of barriers to dispersal that has permitted invasion of nonindigenous species has caused the extinction of more grassland species than any factor except habitat loss (D'Antinio and Vitousek 1992). Nonindigenous species include exotics, which are transported beyond their natural range, and aliens, those that colonize an altered landscape. Introducing nonindigenous species may increase the number of local species, but it reduces integrity, above- and belowground, and also the number of native species, both aquatic and terrestrial.

A more subtle threat to integrity is loss of genetic diversity. Species hybridize along forested corridors that now fragment the Great Plains (Knopf 1986). Human activity, either accidental or deliberate, moves species from one place to another at ever-increasing rates (Knopf 1992). As a result, species that evolved in isolation from one another are forced into contact. In terms of conservation of biological diversity, the loss of six bird subspecies due to the hybridization arising from these forested stepping stones and artificial corridors (Knopf 1986) rivals the loss of three species attributed to fragmentation of the eastern deciduous forest.

These recent forest patches and woody corridors that border rivers on the Great Plains also favor movement of reptiles and mammals from east to west, which contributes to the degradation of the biological diversity and integrity of the Great Plains (Knopf and Scott 1990). In 1842, in eastern Colorado, the explorer John C. Fremont observed that "antelope were tolerably abundant, wolves were seen in great numbers, and buffalo absolutely covered the plains on both sides of the (South Platte) river" but reported no deer (in Nevins 1956). In recent years, deer abundance has increased markedly, particularly that of the eastern white-tailed deer, which may replace the mule deer, known to have occurred on the western plains since before European settlement (Kufeld and Bowden 1995). Hybridization between the two deer is known to occur (Stubblefield et al. 1986).

Nonindigenous species now account for 13% to 30% of prairie species (U.S. Congress, Office of Technology Assessment 1993; C. Freeman, Kansas Natural Heritage Program, Lawrence, personal communication). Increases in distribution and abundance are inevitable without action to prevent them, as evident in the naturalization of Russian-olive trees in the western United States (Olson and Knopf 1986). Russian-olives, which were introduced from Europe in colonial times, range across the Great Plains into the West. In agricultural areas, this species interferes with farming operations; it also hinders management activities on national wildlife refuges, increases degradation of river channels, contributes to declines in river levels, and supplants native riparian tree species. Its adaptability and resistance to control measures, two life-history traits shared with other nonindigenous species, will continue to add to the management concerns associated with the ever-growing number of other nonindigenous plant and animal species.

The approximate action to take after intercommunity management is to identify and retain a set of species and natural processes that sustain communities characteristic of a particular landscape (Chaplin et al. 1996). The Nature Conservancy has identified significant concentrations--legacies--of prairie species that are rare or of declining abundance (Figs. 16-23). Principally these are species that are federally listed or are species of concern that occur within certain communities and ranges of environmental features, including those from prairie wetlands to cottonwood savannahs.

Recommending the restoration of ecological processes in conservation is not new (Leopold 1933). Understanding scale, spatial and temporal, in management is new (Gibson et al. 1993).

Resource and Research Needs

Solutions to the deterioration of grassland resources appear to revolve around a single emerging concept--sustainability--and conform to a proposed strategy. It is important to increase our understanding of the long-term sustainability both of populations of species and of the overall ecosystem. The strategy arises from the following conceptual principles:

* Support public, private, and governmental prairie conservation initiatives as a step toward the long-term goal of grassland sustainability.

The Western Governors' Association's Great Plains Program, the first broad- scale ecosystem management effort in the United States, seeks to demonstrate that economic and environmental interests are served by preventing declines in the numbers of prairie species and their host ecosystems (Clark 1996). The program is built on broad-based science (Johnson and Bouzaher 1996).

* Develop information that recognizes the biological and ecological significance of prairie communities.

The International Institute for Sustainable Development's Sustainable Development for the Great Plains Policy Analysis (Tyrchniewicz and Wilson 1994) links, at an ecosystem level, the well-being of grassland biological resources, particularly soil and water, and society in general. Increasing evidence suggests that dry ecosystems, whether in South America (Mares 1992) or North America (Samson and Knopf 1994), are unusually diverse compared with wet or rain-forest ecosystems. In the face of global warming, the health of planet Earth may depend on grasslands because they are superior carbon sinks compared to forests with similar environmental characteristics (Seastedt and Knapp 1993).

* Identify, inventory, and conserve prairie-limited animals and plants, particularly the large number of plant and invertebrate species.

Sustainability depends on biological diversity to keep all ecological systems, aquatic and terrestrial, functioning and healthy (Lubchenco et al. 1991). Sustainability and biodiversity are two sides of the same coin (Raven 1991). Information and appropriate actions are required to minimize any negative effects on prairie genetic stock and thus on biodiversity, because diverse natural ecosystems help maintain hydrological cycles, regulate climate, absorb and break down pollutants, and contribute to the process of soil formation (Tyrchniewicz and Wilson 1994).

* Evaluate the status of species of concern and other sensitive species, and encourage measures to reverse downward trends in population numbers of prairie species and rare communities.

Putting the maintenance of diversity as a top priority builds ecological knowledge accessible to the public and environmental decision makers (public and private) and provides opportunities to cooperate in conservation of rare species and communities (Chaplin et al. 1996). Sites for tourism and recreation are often identified in the process as well. The net payoff of understanding and displaying diversity is identification of areas of endemism as an essential aid in planning for the conservation of the nation's biodiversity. An urgent need exists to further develop and refine this process on a national and international basis (International Council for Bird Protection 1992).

* Prairie management should mimic the natural disturbance regime to take advantage of preselected traits of prairie species.

The importance of disturbance in shaping grassland communities (Figs. 3 and 10) and ecosystem dynamics is recognized, yet significant questions remain to be answered on the relationship between disturbance and species persistence (Bragg and Steuter 1996; Steinauer and Collins 1996; Weaver et al. 1996). The purpose of conservation is not to conserve species per se but to conserve interactions among species and processes that maintain the health and productivity of communities and ecosystems (Odum 1992).

The Conservation Reserve Program is one of the most popular and successful conservation programs ever implemented by the U.S. Department of Agriculture. By establishing needed grassland, the Conservation Reserve Program has improved game species habitat, prevented loss of topsoil, improved water quality by reducing pesticide and fertilizer runoff, and provided billions of dollars in environmental benefits over the life of the program. The act, however, needs refocusing to be of major conservation benefit to endemic grassland species (Allen 1993).

* Education programs should play an integral role in ensuring conservation of grasslands.

The premise in prairie conservation is that attitudes of individual landowners and the community as a whole play decisive roles in determining the eventual fate of grasslands (Mack 1996). Programs must foster a climate favorable to grassland conservation as an integral component of agricultural land management (Dyson 1996), to the development of covenant agreements to protect remnants (World Wildlife Fund Canada 1988), to cooperative conservation programs between private and neighboring government-managed lands (Bueseler 1996), and to the importance of science-based management (Johnson and Bouzaher 1996).

Almost a half-century has passed since Weaver (1954) noted that the disappearance of a major unit of vegetation--the North American prairie--is an event worth considering. Fortunately, to a growing segment of our society, prairie "looms as large as the universe, as intimate as a village" (Least Heat Moon 1991).