What is extinction?
Biological diversity is lost through the process of extinction, which operates not on species but on populations. (Remember that species are abstractions. As a convenience we classify like organisms under the same species name, but in the real world these organisms live in populations groups of organisms that are more likely to breed with one another than with organisms from other similar groups. The species we define typically are composed of many populations, although if these populations are isolated enough from one another that significant genetic differences arise through evolutionary divergence, taxonomists may argue that they should be recognized as separate species or subspecies witness the debates over classification of different brown bear or wolf populations in the northern hemisphere.

A population disappears when all of the members of that population either die or leave and are not replaced by the birth or immigration of new members. The disappearance of a population is termed a local extinction, or extirpation. When a population goes, it takes with it some of the genetic diversity of the species to which it belongs, including any special adaptations evolved to cope with the local environment. (For example, the grizzly bears that were extirpated from California were probably genetically distinct from the grizzlies that now live in Glacier, which are also probably genetically distinct from the brown bears now living in Alaska.) Population extinctions also erode the habitat or community diversity of an area. When a population is lost, the composition of the community in which it lived changes; if enough populations disappear, or if a keystone species (NH 4.18) is lost, the entire community may "collapse" or become greatly simplified. For example, if a tree species disappears from an area, any insects that depended on those trees for food, any birds and mammals that used those trees to nest in, and any plants that grew only in the shade of those trees may disappear as well.

As populations of a particular species become locally extinct, that species may become endangered. The disappearance of the last remaining population of a species (or perhaps the only population if it is a naturally rare species) is a global extinction.

Extinction is a natural phenomenon. In fact, the fossil record indicates that extinction is the inevitable fate of all species: eventually, changes in the environment cause all species either to adapt and hence evolve into new forms (see NH 4.10, Ehrlich & Ehrlich 1981), or to disappear. But humans are changing the environment faster than species can adapt and are causing species to go extinct at a much higher rate than would normally occur. It is difficult to estimate how many species are going extinct every year as a result of human activities, since we do not even know how many species there are on earth (estimates range from 2 to 30 million), and many species, especially in the highly diverse and highly threatened tropical ecosystems, are probably disappearing before they can be discovered, named, or described. It is undoubtedly safe to say, however, that we are causing at least several thousand global species extinctions and many more local extinctions every year (Wilson 1988).

How do people cause extinctions?
Human activities which cause extinctions can be divided into two categories: direct threats and indirect threats. Direct threats are those activities which result directly in the injury or death of organisms. These include hunting, fishing, trapping, collecting, and harvesting as well as accidental killing (an animal struck by an automobile or train, for instance, or a plant inadvertently trampled). Also included under direct threats are intentional poisoning (for example, leaving poisoned livestock carcasses to kill predators such as coyotes or wolves) and unintentional poisoning (for example, the now-outlawed use of the pesticide DDT which accumulated in the tissues of birds of prey and led to the thinning of their egg shells and, hence, low reproduction rates and endangerment of such species as bald eagles and peregrine falcons).

Indirect threats are those activities which do not directly harm organisms but which modify or destroy their habitats activities which disrupt some or all of the biotic and abiotic components that organisms require in particular combinations for survival. Loss of habitat causes populations to become fewer, smaller, and, therefore, more vulnerable to extinction. Human-caused habitat destruction is by far the single greatest factor contributing to extinctions today. Habitats may be destroyed outright by such activities as building roads (which literally pave the way for further human penetrations into formerly undisturbed areas), clearcutting forests, plowing fields, constructing dams, and developing towns and cities. Other activities may not completely displace a natural habitat, but may modify it so that many species no longer find it a suitable place to live. Habitat modification can result from various kinds of air and water pollution and toxic waste dumping, fire suppression, introduction of exotic species which act upon the native plants and animals as predators, competitors, or pathogens (livestock grazing might be included in this category), changing watershed patterns through stream diversions or wetland dredging and filling, or the intentional or accidental removal of selected native species.

Threats to small populations
Usually it is not a human hand that actually kills that last survivor of a species. Instead both direct and indirect threats fragment and reduce the size of plant and animal populations. These small, isolated populations become increasingly vulnerable to stochastic (random, unpredictable) events which then can lead to their extinction. Such random events are categorized as environmental, demographic, or genetic stochasticity.

Environmental stochasticity includes both normal environmental variations (for example, year-to-year fluctuations in rainfall, temperature, and availability of food) and natural catastrophes (such as fires, floods, earthquakes, and volcanic eruptions). If a species is composed of many large, geographically widespread populations, it is extremely unlikely that a single event such as a flood or fire could wipe them all out at the same time. On the other hand, if only one small population of a species remains, such an event could result in that species' global extinction.

Demographic stochasticity includes random factors affecting births and deaths in a population. For example, it turned out that all of the last six remaining dusky seaside sparrows on the east coast of the United States were males. This chance happening made the natural recovery of this subspecies biologically impossible. If there had been 500 sparrows producing offspring instead of just a handful, it is very unlikely that all of their offspring would have turned out to be the same sex. (Similarly, if you tossed a coin 500 times, you would be very surprised if heads came up every time; but if you tossed a coin only three times, getting heads three times in a row wouldn't be that unlikely.)

Genetic stochasticity refers to the random processes involved in passing genes from one generation to the next. Genes may be lost from a small population when the individuals carrying those genes fail to pass them on to their offspring or die before having offspring. The resulting shift in the genetic makeup of the population is termed genetic drift. When a population loses some of its genetic diversity, it loses some of its ability to adapt to environmental change.

For example, almost all humans have one of four different blood types (A, B, AB, and O) that are coded for by different variations (alleles) of a single gene. Now suppose an ocean liner that is carrying 400 people sinks and only 10 people survive. Among the original 400 passengers, 100 people had each of the four blood types. But of the 10 shipwreck survivors, it turns out that none has Type O blood, and so, by chance, the allele coding for that blood type has been lost from this population.

Next suppose these 10 survivors make it to a deserted island where they remain stranded for many years. Nine of the survivors are men, and one is a woman. This woman marries one of the men and they have children. It turns out that this man and woman both have Type A blood, and therefore their children also have Type A blood. The other men eventually die with no offspring. Hence two more blood types, Type B and Type AB, are lost from the population. In the remaining population on the island, then, only one blood type is represented instead of the four blood types found in the original population of 400 people on the ocean liner. As a result of its reduction in size and random chance, this population has lost some of its genetic diversity.

Another genetic problem faced by small populations is inbreeding depression. The smaller a population gets, the more likely it becomes that any two individuals in the population are related to one another. When closely related individuals mate, their offspring often will be less fit to survive and reproduce for a variety of reasons. Inbreeding depression can therefore cause a small population to become even smaller and more vulnerable to extinction.

How large must a population be to be invulnerable to stochastic events? Shaffer (1981) defines a viable population as one that has a 99% chance of surviving for 1,000 years despite the foreseeable effects of environmental, demographic, and genetic stochasticity. An entire book has been written on the subject of viable populations (Soule 1987), but population viability assessment (PVA) remains an inexact science. Most models used to estimate viable population sizes for particular species either leave out critical variables or require extensive data sets that take years to collect by which time the populations in question may have already gone extinct.

The downward spiral to extinction
Once a population drops below its "viable population" size, it may become irreversibly trapped in a downward spiral towards extinction. For example, suppose a large population is broken into several small, isolated populations as a result of habitat destruction. Some of these populations are wiped out by disease. The remaining populations, now even more isolated, experience higher rates of inbreeding depression, which lowers birth rates and raises death rates. The populations get smaller and, therefore, even more vulnerable to inbreeding, chance demographic events, and environmental catastrophes. Unless humans intervene, this cycle may continue until all the populations are extinct.

But once populations get locked into this downward spiral, it may be too late for us to prevent extinction. Crisis management for example, removing remaining individuals from the wild and attempting captive breeding, as was done with the California condor and the black-footed ferret may or may not succeed, particularly if the ultimate cause of most species endangerment habitat destruction continues unchecked.

How any species reaches this crisis stage depends on the unique biology of that species and the circumstances that lead to its decline. Descriptions below of the history and status of four species in the Crown of the Continent Ecosystem (CCE) illustrate the universalities of the extinction process as outlined above, and some of the peculiarities of the process that must be dealt with on a species-by-species basis.

Grizzly Bear - Ursus arctos horribilis
At the turn of the 19th century probably over 100,000 grizzly bears lived in the western conterminous (lower 48) United States, from the Pacific Coast to the Mississippi River. Today in that area fewer than 1,000 grizzlies survive in isolated pockets which cover only 1% of their former range. The human activities responsible for this decline persist today in varying degrees: agriculture, livestock grazing, timber harvesting, road building, housing development, and mineral, oil, and gas exploration and extraction, all of which destroy grizzly bear habitat; sport hunting and predator control actions which result in direct bear mortalities; outdoor recreation and other activities which displace the solitude-seeking bears from otherwise suitable habitat. (Activities such as hunting and predator control programs have become less common or have been regulated in recent decades, allowing the grizzly bear population in the CCE to recover to its current level.)

Of the six remaining grizzly populations in the lower 48 states, the one in the CCE is the largest with an estimated 440-680 bears (USFWS 1982, Dood et al. 1985). The approximately 22,950 sq km (8860 sq mi) of occupied grizzly habitat in this region is referred to in research and management documents as the Northern Continental Divide Grizzly Bear Ecosystem (NCDGBE). Only 18% of the NCDGBE is Glacier National Park, although with an estimated 200 bears, Glacier may harbor from 30-45% of the entire population. Another 20% of the NCDGBE is designated wilderness (Bob Marshall, Great Bear, Scapegoat, and Mission Mountains) managed by the U.S. Forest Service. About 8% is privately owned and the remainder consists mostly of Forest Service multiple-use areas (see Information Paper 1), with small patches of state and tribal lands (Servheen 1986).

In 1975 the grizzly bear was listed under the Endangered Species Act as a threatened species in the lower 48 states. A Grizzly Bear Recovery Plan was written (USFWS 1982; revised 1990) listing recovered population goals for the NCDGBE, Yellowstone, and Cabinet-Yaak grizzly bear ecosystems. (Recovery goals have not been set for the other three grizzly bear ecosystems North Cascades, Selkirks, and Selway- Bitterroot because not enough is known about the populations or carrying capacities there.) The recovery goal for the NCDGBE includes, over a six-year average, 10 females with cub in Glacier as well as 12 females with cub outside the park, occupying 21 of 23 designated bear management units (BMUs) annually. Mortality must be no more than 4% of the population per year, and of that mortality no more than 30% may be female. Furthermore, successful recovery will include occupancy of the Mission Mountains as part of this ecosystem (USFWS 1990). At that point the grizzly bear could be considered for delisting (removal from the threatened species list).

In the Cabinet-Yaak ecosystem the recovery goal was originally set at 70-90 bears based on Shaffer's (1978) modeling of the effects of demographic stochasticity on grizzly bear population persistence. (His model showed that grizzly populations of this size would have a 95% chance of surviving chance demographic events for 100 years.) This figure did not take into consideration genetic problems in small populations; a population this small might or might not survive short-term demographic events, but would probably suffer long-term loss of evolutionary fitness through genetic drift and inbreeding depression. Recovery goals have been reevaluated in light of these considerations and augmentation programs have begun in the Cabinet-Yaak population. Augmentation involves relocating bears from stable, high-density populations like the one in Glacier, a method that may increase population size and increase gene flow among the populations that have become artificially isolated from one another by human developments.

It can be very difficult to collect field data on the elusive and potentially dangerous grizzlies. After more than 15 years of study, researchers are still struggling to develop practical methods of monitoring grizzly bear population trends so that managers can both assess whether recovery goals are being met and determine whether those goals are valid in the first place. Keating's (1986) use of historic ranger logs to determine historical grizzly bear trends in the park suggested the potential for using standardized ground surveys to monitor the grizzly bear population. Kendall (1992) describes current efforts in Glacier to use such repeated fall ground surveys of bear sign on trails to monitor population trends. Feces and tracks are the most abundant and reliably observed evidence of bears, although distinguishing black bear from grizzly bear feces requires laboratory analysis (Kendall 1992). Additional methods being used to monitor population trends in other parts of the grizzlies' range include radio tracking and camera baiting rigging up bait so that when the bear triggers an infrared sensor, a camera's shutter is released snapping a photo of the bear (Mace & Manley 1991).

The above figures for the grizzly population in the NCDGBE do not include grizzlies living in Canada adjacent to the ecosystem. The bears in Canada are often seen as a "backup" or reservoir of additional bears which will serve as a buffer against genetic deterioration or extinction for the NCDGBE population (Horejsi 1988). Radio-tracking studies do show that bears move freely across the international border (Hamer et al. 1985), but the portion of the population in Alberta, Canada, may in fact be far more threatened than the U.S. portion (Horejsi 1986,1988). Horejsi (1986) estimates that only 16-25 bears use habitat in Waterton Lakes National Park itself in a given year; outside of Waterton on private or provincial land the bears may be subjected to very high mortality from hunting and "predator control" programs advocated by local cattle ranchers. According to Horejsi (1988), at least 59 grizzlies were removed in southern Alberta alone between 1982 and 1988; in particular he cites the Poll Haven area, which borders Waterton, Glacier, and the Blackfeet Reservation and supports around 600 head of cattle, as a "black hole" for bears.

Although some people feel that the Northern Continental Divide grizzly bear population is out of danger, and this population may be considered in the near future for delisting from the threatened species list, it will be prudent to continue monitoring it in order to detect impacts of future human activities in the ecosystem. Grizzly bear habitat continues to be fragmented, encroached upon, and eliminated through housing development, resource extraction, and recreational use. In addition, researchers must stay alert to the possibility of stochastic threats which are by definition unpredictable and unpreventable.

Northern Rocky Mountain Wolf - Canis lupus irremotus
The northern Rocky Mountain wolf is one of 32 subspecies of gray wolf recognized by some taxonomists (Mech 1970). Gray wolves once were distributed throughout most of the northern hemisphere north of 20 degrees N latitude. Due to declines in their natural prey base (ungulates, beaver) and poisoning, trapping, and shooting by humans, wolves have been largely eliminated from the southern half of North America and have declined as well in Canada and Alaska, although somewhere between 61,000 and 63,000 wolves still live in those areas. In the lower 48 United States around 1500-1700 wolves live in Minnesota, 24-32 in Michigan (including Isle Royale), 40 in Wisconsin, less than 15 in Idaho, and about 60 in Montana (Fritts 1992).

Three or four wolf packs were probably living in and around Glacier National Park when the park was created in 1910. By the 1920s predator control programs had extirpated virtually all wolves from Glacier and Waterton Lakes National Parks. Between 1948 and 1956 wolves increased in the area, but from 1957 until the 1980s only an occasional lone wolf was spotted in the region (Singer 1975).

In 1982 a litter of wolf pups was born just north of Glacier in British Columbia and additional litters were born in 1984 and 1985; in 1986 a litter was born in Glacier National Park. These represent the first documented wolf reproduction in or close to the western United States in over 50 years (Ream and Harris 1986). These wolves were probably able to move into this region by following the corridor provided by the Canadian Rockies (Ream 1984). When the first litter was observed in 1982, the closest breeding population was probably 150 miles away in Banff National Park (Ream and Harris 1986).

From 1984 to 1992 the wolf population in the Glacier area has risen from 7 to as high as 30 wolves in 1988 (Ream et al. 1989; K. Kunkel pers. commun.). While the population is small, we would expect to see large fluctuations in numbers of wolves as a result of chance events. Canine parvovirus or distemper, which is known to be carried by some of the wolves in Glacier (Ream et al. 1987), has the potential to wipe out entire litters or packs. Individual wolves may simply not be able to find suitable mates (Ream 1984). Isolation may result in unnaturally high rates of inbreeding, although there is some debate as to the extent to which wolves are adapted to inbreeding (Shields 1981, Theberge 1981).

The main threat to wolf recovery in the United States, however, is human-caused mortality. Wolves are subject to illegal killing, and those known to prey on livestock may be legally killed both in the U.S. and Canada. In 1973 wolves were listed under the federal Endangered Species Act as an endangered species in the U.S. Rockies (USFWS 1987). They cannot be legally hunted or trapped by the public in this region, but they may be legally hunted in parts of Canada. (The alpha female who pioneered the return to Glacier in 1986 was legally shot in Canada in December 1992. She was over ten years old and known to have raised at least seventeen pups.) If human-caused mortalities can be controlled, the wolf may have an excellent chance of establishing a viable population in the northern Rockies in spite of their current low numbers because individuals disperse over large areas and because wolves have a high reproductive potential (the alpha female of a pack may produce a litter of 4-7 pups every year [Mech 1970]), attributes which contribute to genetic interchange and reduce the threat of demographic stochasticity.

Recovery goals for wolves in the northern Rockies have been set at 10 packs (about 100 wolves) in each of three areas: northwestern Montana, central Idaho, and Yellowstone. Because of the Glacier region's link with the Canadian corridor, it has the greatest chance of recovering without human intervention. As of April 1993, Glacier has two packs of wolves in the North Fork of the Flathead River valley. The North Camas Pack, with 12 animals (including 6 pups of 1992), uses territory north of Polebridge; the South Camas Pack, with 13 animals (6 pups of 1992), generally stays in the southern part of the drainage. The Spruce Pack includes 6 animals whose territory is along the U.S.-Canadian border; west of Whitefish is the Murphy Lake Pack with 4-6 members. Elsewhere in Montana, the Nine-Mile pack northwest of Missoula has 4-6 wolves. The only wolf pack east of the Continental Divide in Montana was located for the first time in March 1993. A very large male (122 lbs.) and his mate are being monitored near Augusta. Discovery of this breeding pair was particularly exciting because the male proved to be one of the Wigwam Pack (British Columbia) which had been believed poisoned years earlier. North of the border, the Headwaters Pack of 7-8 animals occupies the upper Flathead valley once home to the Wigwam Pack. East of the Divide, 2-4 animals have been sighted in the Belly River area, usually north of the U.S.-Canadian border. One of the females is a disperser from the North Camas Pack in Glacier National Park.

For several years the University of Montana sponsored the Wolf Ecology Project to study the return of the wolf to northwestern Montana. Since 1992, however, wolf studies have been combined with ongoing work on elk, deer, and moose. The current Wolf-Ungulate study (funded equally by the USFWS and the NPS) focuses on predator-prey relations in the North Fork valley. Thirty female elk, 30 female moose, and 30 female deer have been radio-collared; their movements can be tracked by air and related to those of the wolves collared in the North Fork packs. What effects are wolves having on ungulate numbers and behavior? Conversely, how do ungulates affect wolf numbers and behavior? Three years of data have been collected, but at least another three are necessary before population trends can be determined (K. Kunkel pers. commun.).

Wolf recovery in the Glacier area represents a rare opportunity to observe natural dynamics of recolonization by an extirpated large carnivore and ensuing shifts in community dynamics. The loss of wolves undoubtedly affected many species. Ungulate and coyote populations in this region probably expanded (Carbyn 1987); and some full- and part-time scavengers (crows, ravens, jays, red squirrels, foxes, bobcats, fishers, eagles) that were able to take advantage of wolves' leftovers may have experienced a decline. We might expect to see these trends reverse as wolves move back into the area, but what will actually occur is very difficult to predict since our understanding of enormously complex community dynamics is quite limited. A five-year study of the relations between wolves and mountain lions began in the winter of 1992-1993 with the collaring of 9 mountain lions in the North Fork valley. Researchers on the Wolf-Ungulate and lion studies are working very closely together to learn as much as possible about the complex behaviors and interrelations of these predators. The North Fork is the only place in the lower forty-eight states still home to a full complement of carnivores: wolves, bears, and lions coexist in a virtually complete ecosystem. The North Fork thus presents a unique opportunity for such work on interspecies relations (K. Kunkel pers. commun.).

Wolf recovery ultimately depends more on sociopolitical factors than biological ones. Their successful reestablishment in the northern Rockies depends primarily on the willingness of ranchers to accept some livestock losses and on the willingness of everyone else to either compensate ranchers for those losses or to accept that control measures will have to be taken to relocate or kill depredating wolves, or both (Ream 1984). (See Lopez [1978] and Kellert [1985, 1987] for additional perspectives on human attitudes toward wolves and wolf recovery.) Minnesota, which has 1500-1700 wolves, stands as proof that humans and wolves can in fact live together in harmony; whether they will be able to do so in the northern Rockies remains to be seen.

Rocky Mountain Bighorn Sheep - Ovis canadensis canadensis
Like the grizzly and the gray wolf, bighorn sheep were far more widely distributed prior to the 1800s than they are today. Before 1850 an estimated one-half-to two million sheep belonging to six subspecies of Ovis canadensis inhabited most mountain ranges of the western United States. Between 1850 and 1900 hunting, scabies (a disease caused by mites that were probably introduced by domesticated sheep), competition with livestock for forage, and restriction of winter range by habitat destruction and fire suppression wiped out sheep populations from virtually all readily accessible areas. Today around 28,000 sheep survive in the most remote, rugged, and inaccessible reaches of their former range (Thorne et al. 1985, Weaver 1985). The subspecies O.c. auduboni which inhabited eastern Montana, Nebraska, and the Dakotas is now extinct. Approximately 19,000 Rocky Mountain bighorn sheep are found today in Colorado, Idaho, Wyoming, Montana, Oregon, Utah, Nevada, and New Mexico. Bighorn sheep numbers have practically doubled over the last 30 years (see Buechner 1960) largely as a result of efforts to transplant sheep into unoccupied habitats. One of Montana's approximately 32 native and transplanted populations is located in Glacier National Park and consists of around 250-400 sheep (Thorne et al. 1985, Keating pers. commun.). (In fact, Glacier's bighorn sheep "population" may consist of two or more populations that are more or less isolated from each other; see below.) Several hundred more sheep live in Waterton Lakes National Park (Stelfox 1971).

Bighorn sheep populations go through boom-and-bust cycles (Buechner 1960) which may or may not be naturally caused. Given adequate habitat, their numbers rapidly increase; when a population reaches a high enough density, disease epidemics triggered by malnutrition can kill off large portions of the population very quickly (e.g., Stelfox 1971). Since remaining sheep populations today are so reduced and isolated, however, they may not be as capable of rebounding from such epidemics or they may be altogether eliminated by them.

In Glacier 9 or 10 winter ranges support from 5 to 60 individuals each (Keating 1985, pers. commun.). Any one of these populations may have a high probability of extinction as a result of stochastic events including disease. Because bighorn sheep are poor colonizers and are highly traditional in their use of winter ranges (a learned behavior passed on from one generation to the next), if a population becomes extinct on one of these winter ranges, that habitat may never be rediscovered and recolonized. Some mixing of Glacier's sheep populations with each other and with Waterton populations does occur during the summer months (Keating pers. commun.) and some mixing may also occur during the rut (Geist 1971), which may stimulate gene flow among populations; but mixing also increases the risk of spreading disease which might threaten the entire group of populations (i.e., the metapopulation) of bighorns in the ecosystem.

Some research has been conducted to document the seasonal movement patterns of bighorns in the park, focusing in particular on the Many Glacier population (Keating and Key 1991). However, we are still far from being able to make confident assessments of the long-term viability of bighorn sheep in Glacier.

Westslope Cutthroat Trout - Oncorhynchus (formerly Salmo) clarki lewisi
Although the westslope cutthroat has suffered from habitat degradation and overexploitation, a significant impact on its survival has been the introduction of nonnative species into much if not all of its former range which included western Montana, central and northern Idaho, a small portion of Wyoming, and portions of three Canadian provinces (Liknes & Graham 1988). Stocking of nonnative species to support recreational fisheries was a widespread practice in national parks up until the 1970s. Because of this practice aquatic ecosystems are among the most compromised ecosystems in the national parks today (Marnell 1985).

In Glacier National Park alone an estimated 45-55 million fish and fish eggs were planted between the turn of the century and 1972 (L. Marnell pers. commun.) resulting in the successful establishment of five nonnative salmonid species or subspecies. These include rainbow trout (Oncorhynchus mykiss), brook trout (Salvelinus fontinalis), kokanee (Oncorhynchus nerka), Yellowstone cutthroat trout (Oncorhynchus clarki bouvieri), and grayling (Thymallus articus). In addition, lake whitefish (Coregonus clupeaformis) and lake trout (Salvelinus namaycush), which are not native on the west side of the park, have invaded park waters after being introduced into the lower Flathead River basin (Marnell 1988). Westslope cutthroats sometimes hybridize with rainbow and Yellowstone cutthroat trout. It is estimated that, partly as a result of hybridization, genetically pure westslope cutthroat trout populations are present now in only about 2.5% of their historic range (Liknes & Graham 1988). Lake trout are voracious piscivores (fish-eaters) and kokanee are highly efficient planktivores (plankton-eaters); predation and competition from these two species may also have contributed to declines in westslope cutthroat trout populations, particularly in places like Lake McDonald (Marnell 1988).

Loss of genetic purity in westslope cutthroat trout is of aesthetic and ethical concern, but it is of practical concern as well because hybridization may result in declines in population fitness and the loss of genetic adaptations to the local environment (Marnell 1986). However, extensive genetic research on cutthroat in Glacier has revealed that genetic contamination in the park has been minimal despite the widespread stocking of closely related species and subspecies (Marnell et al. 1987). Genetically pure indigenous populations survive in at least 15 park lakes. Four other lakes have been generally believed to contain transplanted westslope cutthroat populations, but recent evidence derived from sediment coring suggests these populations may be native after all (L. Marnell pers. commun.). One of the reasons for their survival may be that the introduced Yellowstone cutthroat trout lacked genetic adaptations to cope with predatory fishes like bull trout and pathogens and parasites (e.g., tapeworms) that are indigenous to Glacier's waters but not to the waters in which the Yellowstone cutthroat evolved (Marnell et al. 1987).

The pure westslope cutthroat trout populations in Glacier represent the last stronghold for this subspecies; their protection is of the utmost importance both to the park and to other agencies that may in the future use transplants from Glacier to reestablish populations in the trout's former range. Elimination of nonnative species from larger lakes in the park is both technically and economically impractical, but other management practices may help to increase westslope cutthroat populations or at least prevent their decline. Fishing regulations have been carefully formulated to minimize impacts on the cutthroat and maximize the take of nonnatives, particularly the lake trout which account for nearly 50% of the fish caught from Lake McDonald (Marnell 1988, pers. commun.). Imprint planting (placing eggs in a spawning area so that the trout will cue in on that place) may be used in the future to reestablish spawning runs, but problems such as the presence of lake trout would nullify such an effort now (Marnell pers. commun.). (As with bighorn sheep and winter ranges, trout rely on "learned" knowledge, acquired perhaps from olfactory cues, of spawning stream locations; if the use of a spawning area is disrupted by population loss or habitat disturbance, knowledge of that area may be lost from the population.) Reestablishment of spawning populations may provide a secondary benefit by increasing food supplies for bald eagles.

In the longer view, the experimental rehabilitation of small lakes may be a possibility if funding becomes available. This would involve either restoring native populations to lakes or, in some cases, returning lakes to their historically fishless state. The impact of fifty years of an altered aquatic system, however, cannot be quickly changed at the microorganic level (Marnell pers. commun.).

Fish are no longer stocked in most national parks. Glacier National Park officials, however, remain concerned about the possible stocking of nonnative fish species in Flathead and Whitefish Lakes in the Flathead River Basin (see Vashro 1990 for a review of Flathead Lake fisheries management issues). The National Park Service has no direct control over this, but park managers frequently meet with the Montana Department of Fish, Wildlife, and Parks, the U.S. Forest Service, and the Confederated Salish and Kootenai Tribes to discuss options for balancing ecological and economic concerns related to the management of aquatic systems in the region.

Conclusion

Perhaps the biggest lesson to be learned from the case studies above is that there is still a great deal we do not understand about the complex web of interrelationships and processes that support single species, let alone entire ecosystems. The species described above are among the most intensively studied populations in the Crown of the Continent Ecosystem, and yet our ability to predict their longterm viability remains limited at best. For most of the ecosystem's residents, even the large mammals mountain lions, bobcats, lynx, wolverines, mountain goats, moose, etc. we don't even know their population's distributions in the park, let alone how large those populations are or whether they are increasing, declining, or stable.

Since conducting population viability assessments is difficult, costly, and extremely time-consuming, it is unlikely that the technique will be ever be applied to more than a handful of endangered or culturally or economically important species. Finding more practical methods for monitoring trends across the full spectrum of biological diversity is critical if we are to prevent extinctions of other species (see Information Paper 8).

Even in the absence of specific data, however, the second lesson to be learned from examining the extinction process is clear. We should keep populations as large and widely distributed as possible (within natural limits) to minimize vulnerability to stochastic events; and we should not put all our eggs in one basket i.e., we should protect as many separate populations as possible (again, within natural limits) so that if one population does become extinct, we will not have lost the entire species. (It is important to remember that the goal of the National Park Service is to maintain natural populations of native species, not simply to maximize the overall numbers of species and organisms. Efforts to artificially manipulate native populations [e.g., through feeding or relocation programs] or to maintain populations of exotic species are, in general, to be resorted to only when the survival of an endangered species is at stake or an historical landscape is to be preserved [NPS 1988].)

Finally, each of the case studies above illustrates the critical need for cooperation among land managers in the Crown of the Continent Ecosystem (see Information Paper 2). Political boundaries do not hinder the spread of organisms or of human impacts, and they should not artificially restrict the scope of research and management efforts to prevent species extinctions. The interagency efforts on behalf of the species described above may serve as models for future attempts to prevent regional losses of biological diversity.

Author: Karen J. Schmidt.

References

Carbyn, L.N. 1987. Gray wolf and red wolf. Pages 359-376 in M. Novak, G.A. Baker, M.E. Obbard, and B. Malloch, eds., Wild furbearer management and conservation in North America, Ontario Trappers Association, Ministry of Natural Resources, Ontario, Canada.

Tucker, P., ed. 1988. Annotated gray wolf bibliography. U.S. Fish and Wildlife Service.