META-STRATEGY FOR WORLD BEAR CONSERVATION
 
 
(c) 1992 Stephen F. Stringham WildWatch Consulting _________________________________________________________________________ This meta-strategy is an idealized approach. It assumes that adequate data can be gathered and that there is sufficient funding and political will to implement it. In many situations, those conditions aren’t met. Quality of scientific information is limited by funding, time and other practical constraints. And negotiations are often governed less by science than by politics. For example, decision-makers may ignore scientific evidence or minimize opportunities for presenting the evidence. It’s hard to accomplish much in a 2-minute or 1-page window. In theory, human impacts in a reserve might be limited to what the bears (or other wildlife) can sustain. But all too often, conservation measures are limited by what powerful special interests or individuals will tolerate. So real world conservation may deviate far from the meta-strategy proposed here. Nevertheless, this meta-strategy should provide a useful framework for planning and and assessing conservation efforts. Furthermore, although it focuses on bears, many of the component strategies and objectives could apply to a wide range of animals or perhaps plants. EXECUTIVE SUMMARY Beginning with the Ninth International Conference on Bear Research and Management in 1992, workshops have been held periodically to update strategies for bear conservation worldwide. This paper summarizes the key ideas presented and generally agreed to at those workshops. Examples are based on North American bears, but basic principles should apply to bears on other continents and to other taxa. Goals and Objectives are listed in order of priority from the perspective of bear conservation. Impediments to their realization are listed, then strategies for implementation through management, research, and activism are discussed. Goals 1) Maintain or restore permanently viable populations throughout the geographic range of each species, at least in those areas which still provide adequate habitat. 2) Assure that bears remain functional members of the ecosystems they inhabit (as predators, prey, scavengers, seed and nitrogen dispersers, soil disturbers, etc.). 3) Assure that bears are able to maintain natural lifestyles appropriate to their evolutionary history and to modern conditions (conserving [e.g., Panda] bears in captivity is a temporary means of conserving wild populations, not a substitute). 4) Manage bear populations and individuals in appropriate ways to meet the cultural/spiritual needs of people sharing bear habitat (e.g., for bear hunting, watching, and Indigenous cultural practices). Objectives Bear Ecology 1) Preserve bear habitat -- i.e., the ecosystem components and processes -- and the other environmental conditions (e.g., isolation from humans) upon which each bear population depends. 2) Assure that bears remain highly adapted and adaptable. Bear-Human Coexistence 1) Maximize the tolerance of bears for human stressors. 2) Budget impacts rigorously to maximize human benefits per unit of impact without excessive total impact. 3) Assist people to live in harmony with bears. a) Increase reciprocal tolerances between bears and people. b) Educate people on how to minimize dangers and nuisance activities (e.g., management of garbage and livestock). c) Aid industry and the public to identify and utilize technologies that minimize their impacts on bears d) Deal with problem bears (e.g., rehabilitate, relocate, or dispose of them). Impediments Sources of Impact 1) Habitat shrinkage, degradation, and fragmentation. 2) Killing a) Poaching (for personal sport or for profit) b) Defense of self or property; nuisance kills c) Assassination (“The only good bear is a dead bear” viewpoint) d) Inadequately managed legal sport hunting. Weaknesses of Status Quo Management 1) Many critical causes of impact are uncontrollable. 2) Conflicts between bear needs vs. human wants. 3) Conflicts between bureaucratic vs. conservation goals. 4) Scarcity of information on bear responses to stressors. 5) Case-specific failures to use the best available scientific knowledge as the basis for management planning and impact assessment. 6/7) Dependence on unreliable indices of habitat effectiveness/suitability and population status/trend. 8) Management for Minimum Viable Population Size -- Life on the brink of disaster. 9) Failure to consider impacts over a long enough time frame. 10) Ignoring facilitation of future impacts by current impacts. 11) Ignoring cumulative and synergistic impacts. 12) Inadequate flow of information (e.g., within bureaucracies) about impacts and conservation 13) Inadequate expertise/resources for impact assessment & critique. 14) Inadequate statutes and regulations 15) Insufficient enforcement 16) Management “by objectives” divorced from fundamental goals Implementation Management 1) Base management plans on thorough scientific information about bear needs and vulnerabilities. 2) Legalities: management plans should review relevant laws that foster or constrain bear conservation. 3) Only after thoroughly addressing legal then biological requirements and constraints for effective bear conservation, should political and economic constraints be addressed in a management plan. 4) Fifty year test: Assess the consequences of proposed decisions and actions by projecting their effects on bear populations and habitats over the long-term -- e.g., after 20, 50, 100, 250, and 500 year time lags. 5) Manage adaptively: Map out management options to be triggered when population viability or habitat effectiveness fall below specified thresholds -- e.g., thresholds of minimum viability for withstanding: a) unavoidable impacts, b) all the foregoing plus bear viewing, c) all the foregoing plus sport harvest, d) all the foregoing plus habitat degradation, e.g., by logging or mining. 6) Generate feedback of information on success of management. 7) Address uncertainties in data, analysis, and modeling for preserving or restoring population robustness and habitat effectiveness. 8) Integrate bear conservation with ecosystem conservation. 9) Preserve long-term "effectiveness" of habitat. 10) Budget impacts rigorously to a) maximize human benefits per unit of impact, and to b) avoid excessive total impact. 11) Maximize the resistance by bear populations to human impacts. 12) Stimulate recovery of at-risk populations by enhancing habitat effectiveness. 13) Minimize environmental degradation within National Parks and other prime habitat for bears, especially in critical areas for travel, denning, feeding and isolation. 14) Foster low-impact technologies for human activities (e.g., for logging, mining and camping) in bear habitat. 15) Provide financial incentives to individuals and communities to promote bear conservation. 16) Bear "misbehavior": avoidance and treatment. 17) Tailor public education to special interest groups: e.g., hunters, viewers, residents. 18) Generate public support for bears and their management. Increase tolerances for bears by people and vice versa. 19) Enhance governmental capacity to manage bears. 20) Enhance information flows within and between organizations. 21) Frequently re-assess progress towards objectives and the extent to which meeting the objectives is likely to promote meeting goals. As conditions change, reaching goals may require revising objectives and plans for achieving them. Research 1) Analyze and quantify stressors in terms of demographic and habitat impacts, including cumulative and synergistic, direct and indirect, delayed and prolonged effects. 2) Use computers and other means of modeling to simulate the dynamics of populations, habitats (ecosystems), and human societies likely to impact bears. Estimate how population viability and robustness are affected by population size, age-sex structure, habitat effectiveness, gene flow, and other factors. Simulate impacts to bears and to their habitat. Evaluate potential methods/indices for monitoring population status and assessing impacts. 3) Empirically validate models, including those used to assess impacts (e.g., Cumulative Effects Model). 4) Taxonomy: distinguish real vs. phantom subspecies (i.e., genetically- vs. epigenetically-determined differences). 5) Legal evidence: develop methods to reliably and readily identify body parts (e.g., gall bladder) and tissue from at-risk bear species and subspecies so that poached specimens can be identified as such. 6) BioDiversity: preserve the adaptive information in gene pools and memories of bears, and capacities for its propagation. 7) Inbreeding depression: quantify its rate, degree and impacts on population robustness and viability; under what circumstances would inbreeding endanger bears? 8) Assess ecosystem responses to impacts on bears. 9) Literature review, analysis, and synthesis. 10) Standardize data collection and analysis. 11) Accelerate dissemination of information relevant to bear conservation: scientific, technical, legal, cultural, etc. Activism Regulation and Legislation 1) Promote upgrading of the Endangered Species Act and foster its reauthorization. 2) Foster peer review of management plans for at-risk populations and species. Assure that “peers” include members of NGOs and/or universities who have been critical of past governmental plans 3) Promote production of management plans for at-risk populations that adequately address legal, then scientific requirements and constraints, before addressing political and economic constraints (“realities”). 4) Foster ecosystem-level management of at-risk populations and species. 5) Promote stricter limits to ownership, sale, and transport of live and dead bears, in whole or part, where necessary for conservation of the species. 6) Promote compliance with Federal law by all governmental agencies -- Federal (including USFWS, USFS, FERC, BIA, etc.), State, Tribal, county and municipal. 7) Foster refinement of State and Tribal laws to protect bears. 8) Promote adequate consideration of conservation issues in all relevant statutes, regulations, and government policies (private property rights, foreign policy, and economic policy). 9) Foster impact review on all kinds of human activities (e.g., logging, real estate development, bear viewing) which could contribute significantly to impacts on bears. 10) Promote inclusion of an explicit, appropriate definition for each key scientific term (e.g., "habitat") used in each law, regulation, or policy. 11) Enhance protection of "whistle blowers" employed by governments, businesses, and schools (e.g., universities). 12) Promote modified rules for expert testimony to assure that all relevant scientific results are presented objectively and thoroughly before addressing significance of these results for or against the case of any party in a litigation. 13) Promote use of Special Masters (expert judges) to evaluate the factual basis of each case, leaving legal judges to evaluate the legal basis. 14) Tighten specification of the kinds and minimum amounts of information needed and questions to be answered by ERs (Environmental Reviews), EAs (Assessments) and EISs (Impacts Statements) (and their State equivalents) for impacts on bear populations and habitat. 15) Require that research results be accompanied by researcher interpretations of management implications to assure that managers and administrators, along with the public, can benefit from researcher viewpoints and to clarify which interpretations are based on science vs. politics. 16) Periodically review relevant statutes and case law, as well as statements of Legislative intent, legal analyses, and agency regulations. Publish summaries of that information in forms accessible to conservationists. 17) Minimize further environmental degradation within, adjacent to, or surrounding National Parks and other prime habitat for bears. Limit development to areas outside of prime habitat. 18) Place the burden of proof on parties claiming that: a) their activities will not significantly impact at-risk species, or that b) an at-risk species is receiving more protection than it really needs. Enforcement 1) Funding and other support: Increase where necessary, for instance for forensic procedures to identify body parts (e.g., gallbladders) and tissue from at-risk bear species, subspecies, and populations. 2) Critique ERs, EAs, EISs, land management plans, and other Federal, Tribal, State, county and municipal documents relating to bear conservation or to impacts on bears, so as to assure that NEPA, ESA, and other Federal laws are being enforced. Politics 1) ... 2) ... ________________________________________________________________ Keywords: Bear, conservation, Endangered, impact, management, model, rare, risk, strategy, Threatened, Ursidae, Ursus. INTRODUCTION In recent decades, bear biology has progressed rapidly on a tactical level. Hundreds of studies have been done in the field and laboratory, providing a wealth of information about rates of reproduction and survival, growth and maturation, immigration and emigration; population size, age-sex structure, and distribution; food habitats and habitat use, responses to habitat change, physiology and evolution; as well as myriad other subjects (for reviews, see Stringham 1985, LeFranc 1987, Brown 1993). Not only have more studies been funded, but critical new technology such as radio telemetry and mitochondrial-DNA (mt-DNA) analysis have opened doors to discovery. Such tactical information, along with low level strategic advances, have tended to slowed the decline in bear populations worldwide. They have also have set the stage for meta-strategic planning that encompasses all the basic elements for stewardship, covering the full range of viability from those species/populations which are robust enough to withstand high levels of hunter harvest and other human impacts, to those on the brink of extinction. The purposes of a meta-strategy would be to (a) facilitate integration of stewardship across all levels of species/population viability; (b) minimize the likelihood of species/populations declining in viability to Depleted, Threatened or Endangered status; and (c) promote recovery of at-risk species/populations to high viability. In order to achieve those goals, through fostering development of scientifically sounder and more effective conservation strategies, workshops were held in 1992 at both the Missoula, Montana (USA) and Grenoble (France) sessions of the Ninth International Conference on Bear Research and Management. Stephen Stringham and David Mattson chaired the Montana workshop. Stringham also participated in the Grenoble workshop -- which was conducted by the Standing Committee Concerning the Protection of the Brown Bear (Ursus arctos), of the Convention on the Conservation of European Wildlife and Natural Habitats. That workshop updated the Committee's 1988 recommendations for bear conservation. This paper reports recommendations from both sessions without regard to political implications. In addition to editing workshop comments by workshop participants, I have gleaned additional ideas from literature or derived them anew. Citations focus on literature from which ideas were taken or which back points made in the workshops. No attempt was made to keep track of who voiced each idea in the workshops. After the Grenoble conference, I was provided with critiques of the US Fish & Wildlife Service Grizzly Bear Recovery Plan (GBRP) by Mark L. Shaffer (special report to The Wilderness Society) and by Craig M. Pease (pers. comm. to Chris Servheen, Grizzly Bear Recovery Coordinator); these are cited as Shaffer (1992) and Pease (1992). Mattson and Craighead (1994) provided ideas on the ways in which bear conservation could be improved through changing the nature of social interactions among biologists, managers, and other concerned parties. GOALS (In order of priority) 1) Assure the permanent survival of: a) At least one highly viable population of each species. b) At least one viable population of each species in each region of its historic range. c) All extant populations of each species at current or higher levels of viability and size. d) At least one viable population restored to each region from which the species had been extirpated. e) Viable populations in all adequate habitat. "Viability" refers to the a population's capacity to remain viagorous through the foreseeable future -- for instance for one or more centuries. "Vigor" -- the antithesis of vulnerability to impacts -- is a population's (a) capacity to avoid, resist, or tolerate impacts or (b) recovery from impacts -- for instance to grow rapidly when recovering from a catastrophic decline (or when colonizing new habitat). Vigor, in essence, integrates all other demographic parameters, thus providing a summary index to a population's current "health," just as viability is an index to its long-term health. At any given time, a population should be vigorous enough to stabilize or grow in size, vigor, and distribution, despite normal levels of stress from both the natural environment and humanity. The population should also be vigorous enough to survive the stresses of all but the worst catastrophies, and then to rapidly recover from any adverse impact. Catastrophe (e.g., forest fire) could wipe out an entire bear population. There thus need to be enough populations of each species, living in a wide enough variety of environmental conditions, to minimize the risk that all could be wiped out by any foreseeable single catastrophe or string of catastrophes (Fig. 1 [Don't put all your eggs in one basket]). For bears to recover after severe impacts, or to repopulate vacant habitat, they may need to travel between habitats. Travel in turn depends on availability of connecting links. Natural ingress into sparsely-populated habitats may sometimes have to be augmented by importing bears from other populations. Recruits for importation could come from habitat that is "over-populated" and/or from populations with "excessive" viability. Criteria for these thresholds of "over-population" and "excessive viability" have yet to be established. There may be other circumstances where it is preferable to consolidate bears into a single habitat from two or more marginal populations, and then to later restore bears from elsewhere to the newly vacated habitat. 2) Assure that bears remain functional members of the ecosystems they inhabit. a) We have only begun to understand the ecological roles played by bears. So we know little about the ecosystem impacts of severely reducing bear numbers, much less of extirpating bears from an ecosystem. b) Known ecological roles of bears: * Predators: Polar bears feed primarily on ringed seal (__________). North American black and grizzly/brown bears occasionally prey on fish and ungulates, especially fawns. Bear species sometimes also prey on one another. * Scavengers: Bears eat carcasses of animals killed by weather, accidents, disease, parasites, humans, and other predators. * Herbivores: Bears eat a wide variety of vegetation, especially fruiting bodies such as nuts, grain seeds, and fleshy fruit. They also eat a wide variety of herbs. * Seed fertilization and dispersal: Some seeds survive passage through a bear's digestive tract and then are deposited and fertilized at a site distant from their origin. * Nitrogen transport: In habitats with cold wet soils where plants are unable to fix much nitrogen from the air, bear-borne nitrogen from salmon and other marine foods can be a critical source of nitrogen for plant growth and protein formation. * Ground disturbance: In their efforts to obtain floral and faunal foods (e.g., roots, tubers, corms, and rodents), bears often turn over rocks or dig up soil and subsoil. * Intra-generic competition: Bears of one species sometimes kill or displace individuals of another species. In North America, when subadult grizzly or black bears immigrate to habitat where the other species abounds, the dispersers are especially vulnerable; the first species abundant in an area may prevent colonization by the second species (Jonkel 19__). Even when they don't interact directly, bears of one species may limit the availability of food or other resources to the second species. * Intergeneric competition: Bears compete with rodents, birds, ungulates, and other taxa for a variety of foods. * Parasitism: Bears sometimes steal food gathered or killed by other species, as when a bear digs up the nut midden of a squirrel or displaces a puma (Felis concolor) from its kill, then consumes the carcass. Bears themselves are hosts for diseases and parasites such as Trichinosis. 3) Assure that bears are able to maintain natural lifestyles appropriate to their evolutionary history and to modern conditions: There are certain aspects of bear behavior which humans may wish to alter, or may need to alter in order to co-exist with bears. Behaviors to be minimized include reducing aggression towards people, predation on livestock, damage to crops and buildings. The goal of conservation includes such minor changes in bear behavior, but not domestication or shifting bears to live in unnatural habitats or to depend heavily on unnatural food sources such as garbage. 4) Manage bear populations and individuals in appropriate ways to meet the cultural/spiritual needs of people sharing bear habitat: Historically, management of bears in North America initially focused on eliminating harm to people, livestock, crops, and other property. Management later shifted to production of bears for harvest by sport hunters. Those practices have curtailed opportunities for other peoples to benefit from bears. Native Americans, for example, have long considered bears sacred and built components of their cultures around the bear. Destroying bears or manhandling them can markedly impact indigenous cultures. Bears are also increasingly popular as "watchable wildlife." Hunting should not severely impair opportunities for non-consumptive "uses" of bears, and in some cases hunting might best be subordinated to non-consumptive uses -- for instance in National Parks and near the homelands of those Natives who worship bears without exploiting them. Where sport hunting adversely impacts Native cultures, hunting regulations should be adjusted in accord with the Native cultures. OBJECTIVES Bear Ecology 1) Preserve bear habitat--i.e., the ecosystems and specific environmental conditions upon which each bear population depends. (In this sense, an "ecosystem" encompasses enough area and habitat to support at least one thriving, highly viable, population of every species endemic to that type of ecosystem.) The "habitat" in need of protection is not just land with certain physical structures (mountains, valleys, streams, wetlands, etc.) and vegetation (e.g., forests, meadows, bogs). Rather, it is the particular environmental characteristics of such places which are necessary for bears to thrive as individuals and as one or more population. To the extent that bears depend on other animals and plants, and those species in turn depend on yet other species ad infinitum, protection of bears requires protection of the ecosystems. Conversely, ecosystems can benefit in many ways from protection of bears. Given the scarcity of resources for characterizing bear habitat and assessing impacts, it behooves us to focus on the most important and vulnerable kinds and areas of habitat, such as food concentrations (ecocenters), denning areas, and travel corridors. The alternative of giving equal treatment to all kinds of habitat dilutes efforts, sometimes so badly that critical habitat components do not receive adequate attention. We should map, describe, and protect denning areas, ecocenters, travel corridors, and other critical components of habitat. If "effectiveness" of these areas is degraded, for instance by habitat alteration, disturbance, or inaccessibility, then impacts per unit area will be far greater than if impacts were concentrated in less critical habitat. In most of North America, remaining bear habitat is badly fragmented. Some areas, such as the Cabinet-Yaak region of Montana, still retaining a few grizzly bears are completely isolated from other occupied habiat. Among areas still connected by travel corridors, the corridors tend to be narrow and vulnerable to complete severing. For example, the east and west slopes of Montana's Flathead Valley are joined only north of Kalispell and at Evaro, at the south end of the Flathead Reservation; the latter corridor is especially vulnerable and important to grizzly bears, pumas, and other fauna. Areas still joined by travel corridors might be considered as individual "habitats" that are portions of a single "meta-habitat." Likewise, bears living in these habitats could be thought of as belonging to corresponding individual populations within one meta-population. Viability of these individual populations and of the meta-population are already impaired by limited opportunities for travel and gene flow. Habitat effectiveness and population viability would be further reduced if any of these populations is completely isolated. 2) Foster the adaptedness and adaptability of bears. There are at least three modes of adaptation exhibited by animals: (a) natural selection (changes in relative abundances of genotypes and of corresponding phenotypes) and (b) phenotypic plasticity (epigenetics): (b1) morpho-physiological and (b2) behavioral versatility, including learning (Wilson 1975). Each of these capabilities needs to be fostered. Bear-Human Coexistence 1) Maximize bear tolerance for human stressors by maximizing bear population vigor and viability: There will always be limits to the level of impacts that can be borne by any bear population, and from which it will always need protection. Our challenge is to maximize the load that can be carried safely. This is essentially like building up the capacity of an athelete or bracing a bridge (Fig. 3). Just as a powerful athlete might hoist heavier weights than a weak one, so too, a vigorous bear population could tolerate more stress than a marginal population. Conversely, even the most vigorous population has limits and could become overloaded -- devastated (Fig. __) or at least exhausted. If that happens, stress loads may have to be cut back drastically in order to allow recovery. [athletic conditioning analogy] 2) Budget impacts rigorously to maximize human benefits per unit of impact without excessive total impact. Minimize and compensate for low-priority stressors to maximize tolerance of high priority stressors. Tradeoff negotiatons should be shifted from people vs. bears to "this impact" vs. "that impact," or to high-impact vs. low-impact technologies. [REVISE] 3) Assist people to live in harmony with bears: For example: a) Increase tolerances between bears and people for each other. Reciprocal fear between people and bears appears to be a primary cause of mutual avoidance and aggression. Minimize unnecessary/unwarranted fears, for instance through educating both species; bears are of comparable intelligence to apes. b) Educate people on how to minimize dangers and nuisance activities from bears, and to minimize their own impacts on bears: The methods by which people store food, dispose of garbage and carrion, manage livestock and crops, or conduct other activities, can have a large influence on how much they provoke nuisance behavior by bears. How people act in bear habitat can also affect their likelihood of encountering a bear, and whether encounters endanger either themselves or bears. c) Aid industry and the public to identify and utilize technologies that minimize their impacts on bears. Each kind of human activity which could impact bears -- logging, mining, snowmobiling, hiking, etc. -- can be done by a variety of means, methods and mechanisms, some of which have greater impacts on bears than others. People may prefer to select the alternative of greatest profit, convenience or enjoyment to themselves. But where their activities could substantially impact bears, it may be more appropriate to select an alternative that minimizes impacts. d) Deal with problem bears: e.g., rehabilitate, relocate, or dispose of them. IMPEDIMENTS Sources of Impact 1) Habitat Shrinkage, Fragmentation, and Degradation -- Loss of Habitat Effectiveness: Worldwide, a large fraction of former bear habitat has been altered so drastically that bears no longer live there. Likewise, remaining habitat has been degraded, reducing its capacity to support bears. Even otherwise high quality habitat has been so fragmented that few areas are sufficiently large and integrated to support a viable population. Habitat loss is considered to be the single biggest factor causing the demise of bear populations worldwide and within most nations. (e.g., Servheen 1991). (Fig. 5 [habitat frag. floes]). Habitat loss will probably continue worsening over the foreseeable future. Thus, even if there is truth to the claim that the abundance of some bear populations has increased -- e.g., grizzly bears in the Northern Continental Divide Ecosystem of Montana and Alberta -- any gains could easily be reversed by continued loss of habitat effectiveness (Fig. 6). Current carrying capacity can not be maintained if more habitat degradation is allowed before previously degraded habitat has recovered and been repopulated. 2) Killing: a) Poaching (for personal sport or profit): As bears become rarer and opportunities to legally harvest them shrink, some hunters pursue their sport illegally. Worse, some poach for profit. In North America, hunting guides charge many thousands of dollars to help customers kill a grizzly or polar bear. Fees to kill a black bear tend to be lower, but are nevertheless a major motivation for pandering to poachers. Prices in other nations may be even higher. Oriental markets pay ever greater amounts for bear body parts, particularly gall bladders and paws. The value of a single bear can exceed $25,000 (e.g., Servheen 1991). b) Defense of self or property and nuisance kills: Bears are sometimes highly aggressive towards people, trying to prey on humans, to dominate them, or to defend against people. Bears are also lured to prey on livestock, as well as to raid stores of food for livestock or people, corn and other field crops, or garbage. To protect themselves from bear aggression, and to curb "misbehavior," people sometimes kill bears. (Rogers 1987, Servheen 1990). c) Assassination: Especially in areas where economic development or even recreational use of bear habitat is impeded by the presence of bears, or where nuisance activity by bears abounds, these animals are subject to assassination. Two common mottos are "the only good bear is a dead bear," so "shoot, shovel and shutup." d) Weakly managed legal sport hunting: At least three factors limit effective conservation of bear populations managed for harvest: (i) Limited funding and enforcement, (ii) limited incorporation of scientific knowledge on bears into management plans, and (iii) political pressure from some hunters to minimize restrictions on legal harvest, despite conservation needs. Weaknesses of Status Quo Management 1) Many critical stressors are uncontrollable by wildlife stewards: There are many sources and kinds of impact over which bear conservationists have no control, either because such control is blocked by laws (e.g., against "taking" private property) or because minimizing, assessing and compensating for impacts are not mandated by law. Throughout North America and probably other continents, bears depend for habitat on privately-owned as well as publicly-owned lands. In the United States, for example, Federal and State governments have mandates to limit impacts to bear habitat and to bears on public lands. However, governments have little authority to protect habitat and bears on privately-owned land. Accordingly, conservation strategies should be designed to succeed even if all privately-owned land is lost as bear habitat. However, if that cannot be done, and if bear use of private lands is essential, these lands should be purchased or secured by cooperative agreement -- e.g., a conservation easement. Even on public land, there are tight limits to a non-owner's ability to protect habitat and bears. Activities on private property cause ripple effects that degrade public land. There are many conflicting demands from people for use of public lands -- e.g., for commercial exploitation of timber and minerals or for recreation. Critical impacts on bear habitat may originate in distant areas, even in other nations. Global warming, spread of ozone holes, and toxic precipitation (e.g., acid rain) for example, could severely damage bear food sources, and perhaps cause injury to bears. National political and economic policies, such as the North American Free Trade Agreement (NAFTA) could also have major impacts -- accentuating pressures for exploitation of natural resources in bear habitat. Yet these policies are not subject to environmental impact studies under Federal or State laws. The argument is made that the appropriate "place" for doing impact studies is evaluating specific resource extraction projects in bear habitat. However, one political and economic pressures mount, environmental concerns are readily bulldozed aside, especially on private property. Effective conservation may require curbing the political and economic pressures at their inception. 2) Bear needs vs. human wants.--Bear needs are commonly subordinated to competing human uses of bear habitat, even in national forests or parks. Much habitat has thus been lost and remaining areas fragmented and degraded (REF). This happens even where laws (e.g., NEPA and ESA) specifically give priority to bear conservation. Non-compliance is due in part to competing political pressures and self interests -- even by bear conservationists (Mattson & Craighead 1994). 3) Bureaucratic vs. conservation goals.--Some management plans focus on bureaucratic goals rather than on conservation goals. Removal of a bear population or species from "Threatened" or "Endangered" status is called "delisting". Delisting, rather than species preservation, is the stated goal of the GBRP. So far, no proof has been provided that the grizzly bear can persist indefinitely even under the moderate level of protection provided by "Threatened" status, much less without it. 4) Scarcity of quantitative information on bear responses to stressors: Science has clearly established that in bears, as in other mammals, reproductive rate is highest in well-nourished populations (Kirkpatrick 198_). For example, Hatler (1972?) and Rogers (1976, 1983) showed that cub production in the black bear (Ursus americanus) was markedly elevated in years following seasons of peak berry production. Likewise, a number of other studies have shown high cub production and/or survival associated with peak supplies of other natural and human-made foods (Bunnell and Tait 1981, ...). Knowing that better food promotes higher reproduction does not, however, tell us how much a 10% or 20% or X% decrease in berry supply would impair reproduction. Likewise, knowing that adding new logging roads to a forest would tend to displace bears and increase their vulnerability to hunters, does not necessarily tell us how much population vigor and viability would be affected by a 10% or 100% or 1000% increase in road density (miles of road per square mile of habitat or number of square miles of habitat at least one mile distant from any road). Nor do we know how much food supply would have to be increased to compensate for a given (e.g., 500%) increase in road density. Few studies have provided such detailed information, for instance through use of more finely-graded scales of food supply, or through regression plots relating food supply to vital rates. Among the few exceptions are the studies by Jonkel and Cowan (1971), Stringham (1980, 1983, 1985, 1986, 1991a,b), Eiler (1981), and Warburton and Collins (in prep.). 5) There has been little use of existing scientific knowledge as the basis for management planning and impact assessment.--Even on topics thoroughly addressed by scientific studies, little of this information is utilized to design critical Federal and state management plans (e.g., GBRP), computer models (e.g., Cumulative Effects Model -- CEM), or impact assessment techniques (Mattson 1991, Mattson and Craighead 1993, Pease 1993, Shaffer 1993). Relevant information is ignored, misrepresented, misunderstood, or misused. Too little of the information contained in data is extracted. And of the information which is extracted, too little is integrated with other information on bears, much less with information on other species and with general theory. People specializing in one species may not read literature on other bears, much less about other taxa. Weaknesses in plans and models remain unacknowledged by their authors and are often unrecognized by the scientific community, governments, judiciary and public. Federal plans and models are not subjected to unbiased peer review prior to or during utilization. Instead of professional critique, plans and models are exposed only to limited public critique. The critique which is received, even from fellow biologists, it seldom leads to significant revision of management plans. 6) Indices of habitat effectiveness/suitability.--Where scientific data are inadequate to quantify relationships between habitat degradation and declines in bear populations, one must rely on educated guesses. It is educated guesses that underly the USFS indices of "Habitat Effectiveness" (HEI) for grizzly bears (REF) and "Habitat Suitability" (HSI) for black bears (Rogers 1987). Habitat "effectiveness" is a concept potentially encompassing several more familiar ideas such as habitat quality, habitat diversity per unit area, diversity among different areas, accessibility, and disturbance. The USFS approach to assessing impacts on bears is to assign an index of cumulative habitat effectiveness/suitability to a specified impact zone, then estimate how much that index would be altered by a proposed "action" (e.g., logging). The Cumulative Effectiveness index represents the combined contributions by numerous subordinate index/models. Each submodel addresses one measurable habitat variable (e.g., food supply or road density) thought to represent one or more critical habitat characteristics which could be impacted, and thereby significantly affect bear population status. This approach can not provide an absolute measure of habitat effectiveness for bears. But it is allegedly a useful tool for comparing the relative status of a bear population under any of two or more alternative habitat conditions, for instance before vs. after logging, or after logging with method A vs. method B. To estimate impacts by logging, each HEI within the CEM would be evaluated for conditions before vs. after logging. Then values across all the indices would be combined (by multiplying or adding, as the case may be) to arrive at the "cumulative" habitat effectiveness indices before vs. after logging. The difference between before vs. after values is supposedly an index of the "cumulative effects" (impacts) by logging. Unfortunately, the benefits of CE modelling are often outweighed by limitations of the model which can produce seriously misleading results. First, there is no explicit link (Fig. __) between either the individual habitat variables or the cumulative index vs. any empirical measure of population status, such as a vital rate, vigor, viability, or distribution of bears through the habitat. Interpretation of CEM results is thus highly speculative. To illustrate: If logging would reduce the HEI/HSI only 1%, does this really imply that vigor or viability of the bear population would decline only 1% (Fig. __)? Does the ratio between changes in HEI (or HSI) vs. changes in vigor differ according to the level of vigor (Fig. __)? For example, might a 1% change in HEI correspond to a 5% change in vigor when vigor is high, but a 20% change in vigor when vigor is low (Fig.__)? Suppose that logging would allegedly change HEI by 2% in one area and 10% in another area; would the actual impacts on bear population vigor actually be 5-fold higher in the second area; could the actual impacts really be lower in the second area? These are not merely academic questions, but the very crux of determining whether use of HEI (HSI) increases insight about impacts or is seriously misleading. Despite this ignorance about actual habitat-demography relationships, and lack of testing for the models, they are the basis of critical management decisions. One example is the claim that proposed natural gas extraction in the Badger/Two Medicine area of Montana's Northern Continental Divide Ecosystem would have negligible impact on the grizzly bear [REF]) -- a claim at odds with the professional judgements of several bear biologists familiar with such developments in similar grizzly habitat. It would have been appropriate, therefore, to test the HEI model on each of several areas of similar habitat where natural gas development has already occurred and to determine whether the model accurately predicts known responses by grizzly populations (although a double-blind test would be preferable). It would not suffice to show that CEM predicts that devastating habitat changes would extirpate a bear population; there is no need of a computer model to reach that conclusion. The model's value would lie in predicting the amount of impact from seemingly minor to moderate habitat changes. Until the HEI/HSI models are verified empirically in numerous cases, they should be used with great caution and given heavy weight in decisions only if they accord with other sources of knowledge and judgement by professional bear biologists. When potential impacts by habitat change addressed in each submodel of the HEI/HSI are combined, the relative weight given to each submodel may or may not be valid, or may be valid in some ranges (e.g., of food supply or road density) but not in others. This is equivalent to the rhetorical question raised earlier of how much berry supply would have to increase to fully compensate for a 500% increase in road density. Increasing berry supply would tend to have effects opposite to those from increasing road density; but how much change in one variable would balance change in the other variable? Aren't there effects of road density that now improvement in berry supply could counteract? (Fig. __). Another serious flaw of the HE modelling approach is that its use requires large amounts of habitat data which are expensive and time-consuming to collect and which require heavy-duty computer analysis -- making this approach accessible mainly to government agencies and perhaps to industry, stifling critique by other parties, and limiting the number of areas to which it can be applied. If attention were instead focused on the most important habitat types, investments of money, time, and manpower for assessment would be greatly reduced and PC computers might suffice to analyze impact data. CE analysis is commonly applied to small areas of habitat. Indeed, the very complexity of the analysis and the massive amounts of data and computing time required, impair if not preclude use of HE analysis on a large scale. Yet, it is just such large scale analysis that is critical for detecting, assessing, and controlling total impacts on bear populations and meta-populations, habitats and meta-habitats, ecosystems and meta-ecosystems. Fine scale cumulative effects analysis (CEA) needs to be supplemented by medium- and large-scale CEA using empirically tested and validated models. 7) Indices of population status.-- The most reliable way, and perhaps the only reliable way, to monitor status of a bear population, is to monitor population size, age-sex ratios, and age-sex distribution over the habitat; as well as age-sex specific rates of birth, death, and migration. Nevertheless, limitations on resources and difficulties of working with partly nocturnal, sometimes dangerous, large carnivores living in dense vegetation has led some managers to use indices of population status, rather than direct measurements. Common indices include bait-station visits (REF), track counts (REF), and scat counts (REF). The fundamental problem with using any index is to determine how accurately and precisely it is linked to the value indexed -- how tightly they are correlated. Suppose that over five successive years, reproductive rate made the following changes: +20%, -40%, -22%, -3%, +9%. Would an index for reproductive rate change in the same direction, much less to the same degree each year? Over how many years would the reproductive index have to fall before one could reliably conclude that reproductive rate is also falling and accurately estimate the rate and amount of decline? How useful are indices whose accuracy and precision are low? The GBRP, for instance, proposes to monitor grizzly population status by keeping track of (a) the annual number of uniquely identified females with cubs; (b) presence of females with young in a minimum fraction of bear management areas; and (c) number of known human-caused mortalities. Each of these measures is subject to considerable error, and each is insensitive to total rates of reproduction and mortality, respectively. For example, although most mortality of adult grizzlies may well be human-caused (REF), and a high proportion of their carcasses may be found, neither case is true for younger bears -- the age-classes sustaining highest mortality rates (REF). No assessment has been made of how said errors and insensitivity would affect reliability of population monitoring, but severe biases seem likely. Unless biases and imprecision are adequately addressed in the Recovery Plan and its application, they could result in population extirpation. (Pease 1992, Shaffer 1992). 8) Managing for Minimum Viable Population Size: Life on the brink of disaster -- MVPS refers to an estimated minimum number of members necessary for a population to have a specified probability (P%) of surviving at least T years (e.g., 95% for 100 years, Shaffer 1980). Shaffer's diagrams presented MVPS as akin to the brink of a cliff. Above MVPS, increases in populaton size added little to the probability of a population persisting 100 years. Below MVPS, probability of persistance dropped rapidly towards zero as population size declined. Conservationists thus came to articulate their goals in terms of keeping populations at a size of at least MVPS, and preferably safely above that brink. The GBRP, by contrast, aims at keeping average population size just above MVPS. Such an "idiot light" approach is extremely risky. An idiot light is the light on the dashboard of some cars which turns red when engine oil pressure falls too low. Measuring oil supply with a gauge, or by frequently reading the dipstick, can warn a driver that the oil supply is falling, long before it becomes dangerous, when she has many options for curing the problem. But if she doesn't learn of the problem until the idiot light comes on, without warning, she has little margin for error. If she isn't carrying spare oil, or can't obtain it, the engine will be ruined by continued driving. Whether or not an idiot light suffices, or the driver needs a gauge, depends on how quickly oil is being lost and how quickly she can detect the loss and replace it. Managing for MVPS is like trying to keep just enough oil in an engine to stop the idiot light from burning. It guarantees that the population is always on the verge of disaster; someday its luck will fail. If the decision to add oil to an engine were triggered by activation of the idiot light, the practice would be risky enough. But much more dangerous would be delaying addition of oil until several consecutive weeks had passed with the light burning. Engine damage and perhaps destruction would be all but inevitable. Basing management of a bear population on keeping average population size at MVPS could be equally disastrous. The GBRP does not propose using MVPS as the lower limit to population size and seek a higher average. Rather, the GBRP proposes using MVPS as the average -- meaning that actual population size could fall well below MVPS without triggering corrective action. Indeed, population size would be below MVPS roughly half the time, and would have to remain there for several years to trigger protective measures. Bears numbers could suffer severe, perhaps irreversible decline below MVPS before the GBRP's indices of population status would drop low enough to trigger alarm, or at least before managers could assemble enough funds, political support, and habitat to respond and save the population. The above problems with basing management on MVPS are aggravated by the fact that the GBRP's estimate for MVPS is woefully low. It is the figure of 50-90 grizzlies derived by Shaffer (1980) -- who has subsequently disowned them as just a first rough figure. More recent estimates, based on more realistic models, more recent data, and genetic diversity conservation, place MVPS at 400-2000 grizzlies (REF) If one is to base management on MVPS, the cliff-edge of population viability, one might well be guided by the analogy of driving towards a cliff along a road that ends at the brink. Waiting to start protective measures for a species until MVPS is reached is analogous to waiting until a cliff's brink is reached before applying the brakes -- not a game for the faint hearted. Worse yet is the GBRP approach of selecting the brink as the average point where brakes are applied, meaning that roughly half of the time the brakes are not applied until after the brink is passed. To survive driving to cliff edges or to succeed in managing at-risk populations, one must begin applying the brake well before the cliff edge, far enough before the edge to assure stopping before the brink. The faster one is approaching the edge, and the harder it is to slow down, the greater the braking distance required to stop safely. Difficulty stopping would in turn depend on road conditions (e.g., dry and clean vs. wet and icy). Analogously, the faster a bear population is shrinking, and the more difficult that trend is to detect and reverse, the sooner our brakes must be applied. 9) Time frame.--Impact studies tend to be conducted over short time frames, usually starting just before impacts begin and at best continuing through brief monitoring while impacts are underway. They can miss the full scope of impacts whose initial appearance is delayed or whose full effects develop only over long time spans. One example of delayed impact is damage to habitat used only during emergency conditions. Many years may pass between emergencies. But when they occur, lack of access to refugia may decimate or extirpate the population. During the drought of the late 1980's, Arizona black bears aggregated in well-watered but disturbed areas near civilization where they had rarely been seen during previous decades (LeCount, pers. comm.). Another kind of delayed impact is the "ripple effect," beginning at the site of initial disturbance (e.g., clearcutting) and spreading slowly over the landscape. For example, after logging begins, bears may be slow to desert their home ranges, but eventually emigrate to marginal habitat where they fall prey to other human stresses such as being killed as nuisances, by hunters, or by vehicles. For example, the recent apparent influx of grizzly bears on the East Front of Montana's Rocky Mountains may represent not an increase in population size but emigration from severely degraded habitat, especially areas which have been clearcut in national forests and other public or private lands to the west. Even impacts that appear negligible over the span of a few years can be magnified over decades or centuries, possibly causing a serious decline in population viability. This is roughly analogous to growth of ruts left by skidding logs up a steep slope. After months or years of water erosion, ruts that were initially less than a foot deep and a few feet wide can grow into major gullies. 10) Facilitated and indirect impacts.--Impacts from one project can in turn facilitate future impacts from other sources -- a fact seldom addressed in impact assessments. For example, cutting a road to allow timber harvest may increase access by hunters and hikers who kill or at least disturb bears. Presence of the road could also increase the profitability and thus probability of mining or real estate development in local bear habitat. Construction of roads and clear cutting could also promote erosion of soil that degrades streams and eliminates supplies of fish upon which might bears depend. Together, such "secondary" or "indirect" effects of logging could greatly compound its direct impacts on bears. 11) Cumulative and synergistic impacts: Impacts often accumulate in such small increments that the individual impacts are judged negligible and permitted, despite their devastating total effects. America's National Environmental Policy Act (NEPA) (Sec. 1508.7) describes cumulative impact in the following terms: "[It] is the impact on the environment which results from the incremental impact of [a specified] action when added to other past, present, and reasonably foreseeable future actions, regardless of what agency (Federal or non-Federal) or person undertakes such other actions. Cumulative impacts can result from individually minor but collectively significant actions taking place over a period of time." The CEA approach focuses on all of a project's impacts on bear habitat within the designated "impact zone" or "affected environment." CEA also addresses all mortality factors associated with a given project (e.g., timber sale). Only those "cumulative effects" are addressed. That approach is important; but it should not be a substitute for analyzing the cumulative and synergistic effects by all local and regional impacts on the bear population and on the ecosystem's capability to support that population. The significance of seemingly minor impacts may be detectable only from a long-term perspective, as they combine with other impacts additively or synergistically and are magnified. "Synergism" refers to the total impact from two our more kinds or sources being greater than the sum of the individual impacts. Synergistic impacts accumulate by non-linear addition or by multiplication (Fig. __). As an example of a non-linear accumulation: if there are three travel corridors between two regions of habitat, loss of one or two corridors might have moderate effects, whereas loss of all three could be devastating (Fig. __). What is lost is often less critical than what is left. As redundancy in habitat features is eliminated, habitat effectiveness and population viability drop geometrically. Thus, now that little grizzly habitat is left South of Canada and most of that is fragmented and degraded, additional habitat changes of kinds and degrees that had tolerable impacts in the past may have devastating impacts now. Better ways are needed to assess the significance of individual impacts so that their cumulative and synergistic effects are predictable and kept within tolerable limits. 12) Poor information flow.--Decision-makers base decisions on information and other influences from numerous sources. These sources include their superiors, peers, and subordinates, some of whom may be biologists; personnel from other governmental agencies, consultants, politicians, lobbyists, and the general public; legal statutes and regulations; scientific reports and publications. Some relevant information does not reach them, arrives in distorted form, or is too demanding to be easily read and understood. Alternative information or influences, such as lobbying, may thus have disproportionate sway. (Mattson and Craighead 1994). 13) Scarcity of Expertise and Resources.--Impact assessment is expensive, often demanding resources that only government agencies and large corporations can afford. This limits the ability of conservationists to make their own impact assessments and to critique evidence allegedly showing that impacts would be tolerable. 14) Inadequate Statutes and Regulations.--Regulations affording protection to bears are often vague or misdirected. Some laws fail to address key factors impacting bears, or address them inadequately or inappropriately. For example, Vermont's Act 250 protects habitat from "destruction," a term normally interpreted as destruction of gross physical features such as draining a wetland, cutting a grove of trees in which deer shelter during winter, or cutting trees that provide nuts or fruit to deer or bears. It does not explicitly protect less obvious but equally critical characteristics essential to habitat effectiveness, such as access for bears or isolation from disturbance by people. Key words such as "habitat" need to be defined appropriately (e.g., in terms of effectiveness rather than gross physical features). Furthermore, definitions should be clear enough that there is little ambiguity about the types, quality, and quantity of scientific evidence needed to assess impacts. One might also question the utility of laws that do not clearly specify burden-of-proof. For example, with each law, is the burden to show that bears would/would not be significantly impacted; or that a species is/isn't ready for delisting? 15) Insufficient Enforcement.--Compliance with laws is impaired by the common lack of awareness by agency personnel of legal requirements (Mattson and Craighead 1994). Penalties for offenders and funding for enforcement agencies may be inadequate. For example, although NEPA mandates that EAs and especially EISs reveal the environmental impacts (by human activities over which the Act has jurisdiction), in fact EAs/EISs are commonly used to mask impacts rather than to reveal them. [... The ESA mandates ... yet ....] [Self defense is invoked all to easily to excuse killing grizzlies. What penalties for provoking attack, if only through carelessness?] IMPLEMENTATION The aforementioned impediments can be overcome through enhanced strategies of management, research, and activism. Management 1) Legal requirements and constraints.--Bear management planning and impact assessments should comply with Federal, State, and Tribal laws. To assure this, one of the first subjects addressed in a management plan should be legal requirements and constraints for conservation of the species in question. Regarding management of at-risk populations of bears, appropriate sections should be cited or quoted from NEPA, the Endangered Species Act, and the Clean Water Act. For example, if NEPA guidelines for addressing cumulative effects were quoted in an EIS section on CEM, it would immediately be obvious that CEM does not suffice to adequately address cumulative effects. 2) Improve the scientific basis for management.-- a) Base each management plan and EA/EIS on thorough incorporation of relevant scientific data, theory, modelling and reasoning. For instance, where appropriate, determine population size, age-sex structure, and geographic distributions. Determine age-sex specific rates of reproduction, survival, immigration, and emigration. Identify mean rates and measures of uncertainty for each mean value. Variance measures precision, the variability of data (e.g., around the mean) within a given data set. Accuracy (e.g., of the mean) should also be addressed. 2b) Identify ranges of environmental, demographic, social, and individual (e.g., health) conditions adequate to assure vigor and viability of each bear population. For example, what are the limits to tolerance for each kind of environmental stress? How do these limits change circumstantially -- for instance relative to vigor and size of the bear population or effectiveness and size of its habitat? c) Subject management plans, analyses, and models to peer review. Assure that peer review is professional in quality and objectivity. Thus far, peer review has revealed no reason to expect that grizzly bears would be adequately protected by USFS CEA and USFWS GBRP approaches to assessing impacts on habitat and population status. 3) Political and economic constraints.--Once legal compliance and scientific soundness of a plan are assured: (a) Discuss political and economic constraints on the application of each alternative conservation approach, indicating which approaches are both adequate to conserve bears and acceptable to the public. If no approach is both effective and popular, explain why. When comparing alternatives, utilize a NEPA EIS type of structure where appropriate. (b) From the spectrum of effective, acceptable alternative approaches, select one as preferable and justify the choice. (c) At least for the preferred alternative and a variety of less popular but effective alternatives, specify management objectives and methods in practical, tactical terms (e.g., sample sizes, frequency of sampling, methods of sampling). In accordance with the above guidelines, revise the GBRP (cf. Peace 1992, Shaffer 1992, Mattson and Craighead 1994). 4) Fifty year test: Evaluate the consequences of proposed decisions and actions by projecting their impacts on bear populations and habitats over the long-term -- e.g., over 10, 20, 50, 100, 250, and 500 year time lags. Two ways of evaluating long-term impacts on bears are by (a) estimating what population size or vigor would be after each specific time-lag, or by (b) estimating the probability that size or vigor would still exceed certain thresholds after said lags. The ultimate criterion of impact on a population should be the amount of decrease in population viability over short or long time spans. 5) Manage adaptively.--Management plans should map out tactical changes in management to be triggered by specific changes in each bear population or in its habitat. For example, protective measures might be linked to thresholds of viability or habitat effectiveness. Suppose that there is some maximum level of viability (VM) that any species or population could attain, such as a 99% probability of still being vigorous 500 years later. Threshold levels for triggering alternative stewardship responses might be set at specified percentages of maximum viability (Fig. __). Using an hypothetical example for explanatory purposes only: * 90-100% Laissez faire use of bear habitat and of bears: * 75-90% Light restrictions on human impacts on bears and their habitat. Regulated sport hunting, based on maximum sustainable yield. Measures taken to minimize attractants (e.g., garbage) leading to bear mortality. * 50-75% Moderate restrictions on human impacts on bears and their habitat. Hunting season shortened or a quota system established. Fines for poor handling of garbage. * 37-50% Threatened status: Sport hunting forbidden, natural resource exploitation in bear habitat limited to certain methods and times of year. Key feeding or denning areas closed to human access. * 25-37% Endangered status: Roads leading into bear habitat closed and exploitation of natural resources forbidden in that habitat; alternative food sources developed and plans established to import bears from other populations. * 1-25% Emergency status: Augment population by importing bears from other populations. Large areas of habitat closed to human access. 6) Manage so as to generate feedback of critical information on management success: Monitor to detect key changes in viability, other demographic parameters, habitat effectiveness, etc. as necessary to assess population status and likely changes in status. Monitor factors likely to cause changes in status so that changes which do occur can be understood and potential changes anticipated and avoided, if desired. Utilize management techniques designed to generate information about the status of bear populations. For example, experiment with changes in harvest intensity or in habitat effectiveness to test hypotheses about how bear distribution would be affected. On the basis of those data and information from the scientific literature, periodically refine management plans and models. 7) Uncertain data and loose population control: Uncertainty should be a key parameter of each variable (e.g., birth rate) addressed in management plans and models. There is uncertainty about both realism of any model and of data processed by the model. For example, the confidence bounds around an estimated birth rate (e.g., 1 cub per adult female per year) could be +25%. Fig. __ plotted a schematic relationship between level of protection needed relative to estimated viability -- tacitly assuming that viability could be measured exactly. In fact, all measurements are imprecise and inaccurate to some degree. Inexact, outdated information on population status and threats, along with impaired corrective actions ("too little, too late"), could allow a population to suffer major decline before the decline can be detected and reversed. Such loose control of at-risk populations is likely to lead to their demise unless compensated by very tight restrictions on human impacts. Conversely, the more accurate, precise, and timely one's information, and the more quickly and effectively one can correct declines in population status, the less restriction would be needed on human impacts to achieve a given degree of safety. That is, the tighter one's control of a population, the lower the apparent level of viability could be allowed to fall before triggering tighter restrictions (Fig. __). Were protection thresholds based on "average" measures of viability, then the same apparent levels of viability would trigger the same levels of protection, irrespective of whether control of the population were tight or loose -- despite the greater danger inherent with loose control. To offset these dangers, it is preferable to trigger protective actions not by averages, but by lower confidence bounds (e.g., 95% or 99%) (Confidence bounds should be determined for each set of measurements.) That would minimize dangers arising from loose control. Note that confidence bounds should not be based solely on sample variance (imprecision) as in typical statistical analysis. Inaccuracy also has to be taken into account. Management is adaptive when the level of protection used varies according to circumstantial need. For instance, making protection thresholds proportional to risk. A familiar example from daily life is adjustment of braking distance and tire traction according to road conditions. The heavier the vehicle and the poorer traction on the road surface, the farther from a stop sign the brakes should be applied. A car can stop more quickly when pavement is dry and clean than when it is cover with wet ice. So too, a car can stop on ice more quickly with high traction (e.g., studded) tires, and especially with chained tires, than with summer tires (Table 1). If road conditions are known, tire traction and stopping distance can be adjusted appropriately. One would know when to replace summer tires with winter tires and when to use chains, as well as how soon to start braking when approaching a stop sign. By contrast, if one cannot predict road conditions approaching the stop sign, one would have to assume the worst and brake as though there were wet ice, in order to be sure of stopping. Whether one is willing to risk not always stopping soon enough normally depends on the consequences. If errors involve negligible penalty, one is likely to take the risk. But if any error could be fatal, great caution is warranted (Fig. 7). Adaptive mangement of a bear population would work analogously. The stopping distance plot (Fig. __) assumes a linear relationship between breaking action and stopping distance. By contrast, the relationship between viability and protection level needed is probably geometric (Fig. __) -- an assumption based on the fact that vulnerability of a population increases geometrically as viability declines (REF). Because loose control of viability requires a much wider margin of error, and thus tighter restrictions on human impact, it behooves producers of impacts and resource managers to support tight control. Conservationists should develop quantitative measures of how (a) data uncertainty declines as sample size increases, and (b) how uncertainty varies according to the method of data collection, and the amount of time elapsed between occurrence of an event and measurement of its impact on bears. Conservation strategies should be capable of preserving bears even during worst case scenarios of model errors, biased and incomplete data, catastrophic impacts, delayed impacts, and delayed corrective conservation efforts. The GBRP does not even approach meeting this challenge. The GBRP and CEA virtually ignore uncertainty and its effects on vulnerability of bear populations. If, for example, the CEM estimated that development of a natural gas well would reduce habitat effectiveness only 0.3%, what would be the confidence bounds around that point estimate (e.g., 0% - 30% at 95% CB)? Again: the less certain one's information and the looser one's control of population status, the broader the margins of error needed to conserve bears, and thus the tighter the restrictions on human activities required for any given level of safety. 8) Ecosystem Conservation: Integrate conservation of bears with conservation of ecosystems. 9) Long-Term Effectiveness of Habitat: When assessing the "effectiveness" or value of habitat, one key criterion should be its long-term protection status -- better protection warranting higher ratings. Management strategies should acknowledge the ephemeral nature of bear habitat on privately owned lands. Assessment should be done to determine whether existing public lands would suffice to support a highly viable population of each bear species. If not, then (a) the goal of maintaining such a population should be replaced with a less ambitious goal (e.g., lower viability); or (b) compensatory protective measures should be introduced (e.g., complete moratorium on logging or mineral development on the public lands); or (c) measures should be implemented to assure protection of key private lands. Conversely, if current public lands could suffice to support a population at an acceptable level of viability, damaging uses of that publicly-owned habitat would have to be governed accordingly. Where public lands are impacted, it may be necessary to compensate by securing ownership or conservation easements on private land. The amount of compensation needed for degradation of public land would have to be judged according to uncertainties in measuring impact so that an adequate margin of error is maintained. 10) Budget Impacts: 11) Maximize Tolerances: 12) Stimulate recovery of endangered populations.-- Eliminate significant but unnecessary impacts. Budget impacts to keep their limits within bear tolerances. Enhance food sources, as well as habitat quality and integrity. In areas where habitat has been degraded, but can be restored, do so -- for instance, by closing unnecessary roads and fostering growth of plants which provide food and cover for bears. Emphasize plants natural to the region which are resistant to expected climate change (e.g., global warming, increased acid rain) and which produce important forage for bears of kinds scarce in the region. In areas where natural food supply often fails, driving bears to raid agricultural lands (e.g., fields of corn or other grain, orchards, beehives, etc.), buy or lease these specific lands and crops; manage them as food sources for bears. Where wetland foods are scarce but important, foster protection of wetlands and creation of new ones managed for the benefit of bears. For example, one might close logging roads and restrict human access in several areas after logging is finished, then promote growth of berry bushes in the logged area. Such compensatory measures might enable a bear population to tolerate new impacts, from a higher priority source (e.g., logging in a new area), that would otherwise have severe impact. Land managers and other decision-makers should balance impacts from various sources to assure that impacts are neither wasteful nor excessive. [Fig.__: Straw that broke the camel's back]. 13) Minimize environmental degradation within National Parks and other prime habitat for bears, especially in critical habitat types such as denning areas, travel corridors and food concentrations. 14) Foster low-impact technologies: Where a human activity (e.g., logging) could degrade bear habitat, select modes of the activity that minimize degradation. Promote development and application of technologies for human activities in bear habitat that minimize the impacts of these activities (e.g., logging, mineral extraction, road construction, snowmobiling). For instance, to minimize impacts of logging, harvest might be done while bears are denned, hibernating, rather than during other seasons. Impacts might be further reduced by using horses rather than large skidders, and then by using pickups rather than much larger trucks ("tractors") to haul logs out of the forest to a shipping site. It is possible that by choosing appropriate means of commercial and public activities, these activities need not be seriously curtailed within bear habitat, except in areas most critical to bears. 15) Financial incentives.--Establish and maintain organizations and funds to compensate people for losses from bears harming livestock, crops, or other property. Simplify and accelerate administrative procedures for payment. Perhaps increase payment to above-market prices for livestock and crops -- recognizing that only a fraction of the depredation can be documented and only at considerable trouble to the owner. Where depredation is too frequent to be compensated on a case-by-case basis, perhaps guarantee purchase of the surviving herd/crop at market time at a set percentage above (e.g., 110%) fall market price. For example, purchase of a sheep flock might be guaranteed in spring, after lambing. The price paid could be determined by the number of sheep alive at the spring census, irrespective of how many survive until market time. A bonus could be paid to the rancher for high survivorship in the flock, maintaining the rancher's incentive for good husbandry. A bonus might also be paid for reducing grazing pressure on habitats where sheep consume or destroy plants important to bears. Cheating would be a temptation. A few ranchers might steal their own livestock. If this seems a significant problem, counter measures should be considered. Promote products (e.g., livestock and crops; arts and crafts; wild berry jams, candies, and syrup) from communities that enhance viability of a bear population. Establish specific contracts for economic development aid in exchange for specific measures to favor bears. Continuation of benefits after the first year or two might be made contingent on evidence that the bears are actually benefiting. 16) Bear "Misbehavior:" Avoidance and Treatment.--Assure that garbage, bee hives, orchards, and other attractants are managed so as not to provoke nuisance activities. Exploit the learning ability of bears to curb nuisance behavior; use sophisticated behavioral modification techniques. Establish "rehabilitation" centers to break ursine "bad habits" so that, after "graduation", they can be returned to the wild. Bears might be "trained" to understand that they need fear only (a) people to whom they are aggressive or (b) people present in specialized circumstances -- e.g., those wearing bright orange clothing during fall, outside the boundaries of protected areas. There is reason to suspect that bears respond to being hunted by becoming more aggressive towards people and in so doing fail to distinguish between hunters vs. non-hunters. To maximize bear tolerance for people, it may be necessary to curb hunting in selected areas over selected periods. 17) Education.--Tailor educational programs on bears to groups such as hunters, farmers, ranchers, loggers, miners, hikers, etc. -- groups facing specialized challenges for coexistence with bears. Conduct courses in the classroom and field, and over telemedia. Address economic benefits from bears and bear conservation; bear needs and vulnerabilities; etc. 18) Public support.--Assure that regulations and penalties have the support of the public and are enforced. Where support is lacking, enhance it to adequate levels. 19) Enhance governmental capacity of manage bears.--Assure that each state/nation has adequate bear management/ conservation plans, monitoring data, enforcement capability and funding needed to maintain or restore population size and viability, as well as habitat effectiveness. Hunting should be managed so as not to (a) seriously impair opportunities to appreciate bears at reasonably close range, or to (b) seriously impede bears from fulfilling their ecological roles. Develop a central office within each nation to lead and coordinate its bear conservation. 20) Information flow.--Assure that decision-makers, managers, and researchers receive thorough, unbiased, accurate information about the status of bear populations and habitats, threats to bears and habitats, and other relevant matters (see Mattson and Craighead 1994 for a detailed discussion). Research 1) Analyze and quantify impacts in terms of demographic and habitat responses, from the following perspectives: a) individual b) direct: (e.g., bears killed by hunters) c) indirect (e.g., reduced food supply lowers reproduction) d) cumulative (impacts add up spatially and temporally) e) synergistic (total impact greater than the sum of component impacts). f) delayed (not all impacts begin immediately) g) prolonged: (some impacts continue for many years). Determine more exactly how bear behavior and population viability respond to a wide variety of habitat changes and other human actions so that one can reliably predict impacts from future actions and the effectiveness of proposed mitigation. Types of impact of particular concern include development of oil and gas reserves, coal and other mineral resources, geothermal energy, utility right-of-ways, real estate, grazing, off-road vehicular operation, outfitting, hunting (especially with dogs or baits), hikers, photographers, and roads. For example, assess impacts on bears by different harvest management plans and hunting techniques (dogs, baits, etc.). Address bear vital rates, physiological stress, disruption of feeding during fall hyperphagia, foraging during the day vs. night, foraging efficiency, and abandonment of habitat. Assure that all human activities and influences likely to impair bear population viability and habitat effectiveness are addressed in strategic plans, models, and impact analyses. Impact assessment needs to be based on long-term views of consequences to bears and other components of bear ecosystems. Delayed impacts can be important. We need to measure time lags between onset of human activities (e.g., logging) and any resulting declines in habitat effectiveness or bear population vigor/viability. Estimate the extent to which current status of each bear population reflects the full impact of past human activities (e.g., destruction of habitat which is critical only during particular weather conditions such as prolonged drought). 2) Simulate population dynamics and habitat dynamics using computers and other forms of modeling. Estimate how population vigor and viability are affected by population size, age-sex structure, habitat effectiveness, gene flow, and other factors. Simulate impacts to bears and their habitat. Evaluate potential methods/indices for monitoring population status and assessing impacts. Simulate changes in population size, age-sex structure, and vital rates, then assess their potential effects on population vigor and viability. For instance, if X bears are killed and reproduction reduced by Y% for t years, how quickly could the population recover thereafter? Link demographic models to habitat effectiveness models, then assess demographic responses to degradation of habitat effectiveness. Assess potentials for cumulative and synergistic, direct and indirect, delayed and prolonged impacts. Estimate the amount of habitat, kinds of habitat, etc. that would be needed to support the population at high viability. Estimate uncertainties in (a) measuring each parameter, (b) making calculations with the data, and (c) modelling population dynamics and habitat change. On that basis, estimate corresponding margins of error needed to safeguard the population. For example, to assure survival of a population of at least 400 grizzly bears in the Northern Rockies, managers might need to aim at maintaining an average of at least 1000 grizzlies. As noted earlier with regard to population viability estimates: the less precise and accurate the means used to monitor a variable (e.g., cub production), the wider the margin of error needed in applying that information to conservation of at-risk populations. 3) Empirically validate models including those used to assess impacts (e.g., Cumulative Effects Model). Assess the validity of each model for predicting impacts from proposed habitat change (e.g., natural gas extraction in the northern Rocky Mountains). Do this by testing the model on comparable cases where impacts have already occurred. Compare model outputs/predictions with empirical observations and measurements -- using double-blind procedures where appropriate and feasible. Again: It is not enough to show that a model can predict decimation of a bear population by major habitat degradation or loss; no model is needed in such cases. The appropriate test is whether the model can predict the degrees of decline in population vigor and viability due to low and moderate levels of habitat degradation or loss. Testing should include cases where seemingly minor habitat change were followed by severe harm to bears. Discard or revise invalid models. Validation of indices would also require documentation of (a) causal linkages between each index and the variable to which it is supposed to correspond, as well as (b) linkages between each index and other factors influencing it. Then demonstration should be made that those other influences do not significantly bias results through either imprecision or inaccuracy. 4) Taxonomy.--A large number of alleged subspecies have been identified in American bears (e.g., Merriam 1918, 1919; Hall 1984). Re-evaluate the taxonomy to distinguish "phantom" subspecies (Geist 1991) from actual ones whose differences represent adaptations to local environments or genetic drift (cf. Goldman et al. 1989, Shields and Kocher 1991, O'Brien and Mayr 1991). Distinguish hybrids which are less adapted than parent subspecies from those which are better adapted; it is more important to conserve invigorated hybrids than depressed hybrids. 5) Legal evidence.--Identify phenotypic traits (e.g., dental and skull morphology) and genotypic traits (e.g., mt-DNA) by which biologists, managers, and law enforcement officers can reliably, quickly and inexpensively distinguish body parts from each endangered bear species and subspecies (Geist 1992). 6) BioDiversity.--The essential importance of genotypic and phenotypic (bio)diversity is that it sometimes enhances adapteness and adaptability. The relative fitness of alternative phenotypes varies along with environmental conditions, both spatially and temporally. Phenotypic diversity is in turn a product of genotypic diversity or of phenotypic plasticity. According to classical theory, if a population exhibits several genotypes, each may be superior under a different set of environmental conditions, or superiority may vary between only certain genotypes, whereas others are always inferior. Eliminating or minimizing deleterious genotypes -- reducing genetic diversity -- could enhance average fitness of the population relative to existing conditions. On the other hand, as the environment changes, formerly inferior genotypes may now confer the highest fitness. (REF) One can seldom predict where, when or how currently deleterious genotypes might prove beneficial. That constrains our ability to assess the adaptive value of genetic diversity and to understand its conservation implications. For example, if a population has a distinctive pattern of mt-DNA, is this difference the result of genetic drift or natural selection? Is the population genetically adapted to local conditions? What are the probabilities of existing mutations being favorable, and of spreading within or among populations? If rare alleles are unlikely to be beneficial or to spread, is preserving them futile? Perhaps more important to the conservation of bears is phenotypical plasticity. Within many taxa, there are clines of body size and brain size -- body and brain size tending to increase together. Likewise, there are size clines between taxa, for instance between animal Classes and Orders (REF). As is commonly discussed in terms of "r- vs. K-selection," rates of metabolism, body growth, population growth, reproduction and mortality tend to be highest in small-bodied species, declining in progressively larger species (REF). Species of small body size tend to be more abundant, reproduce more quickly, and produce more offspring; they also tend to have higher rates of mortality, population growth, and body growth. These are so-called r-selected species. Their dependence on adaptation through natural selection, more than through versatility (Wilson 1975), puts a premium on maintaining high genetic diversity within and between populations. Although many alleles may be deleterious under current conditions, this "genetic load" is maintained by mechanisms such as heterozygous masking of deleterious recessive genes. Once conditions change, formerly deleterious alleles may in fact become the most adaptive. Species with larger bodies are better equipped than those with small bodies to adapt physiologically -- e.g., acclimation. Those with larger brains are probably better able to adapt through behavioral versatility, especially through learning. The highest degrees of physiological and behavioral adaptability tend to be found in so-called K-selected species, those with sparse populations, low rates of reproduction, mortality, population growth, and body growth. These traits minimize the advantages of maintaining genetic load. Most environmental change can be adapted to without alteration of the genotype. There may be much less difference in the fitness of two underlying genotypes than there is between the phenotypes if any phenotypic differences result less from genetic differences than from plasticity. Within any taxon, predators tend to be less abundant than prey species, and large-bodied, large-brained predators tend to be least abundant of all. So it is in these taxa that one would expect to find least genetic load, and thus least genetic diversity (REF: Allendorf ??) According to that hypothesis, bears should depend less than most other mammals on natural selection for adaptation, and might well have among the lowest levels of genetic diversity. Conversely, they should depend more heavily than most carnivores on learning, through both individual experience and tradition. We should determine how important learning (e.g., of foraging sites and techniques) is to local adaptation, and how much genetic vs. "biocultural" diversity each contributes to population viability. Is saving rare genotypes less critical to viability than preserving biocultural heritage? Might both sometimes be less important than maintaining population size and age-sex ratios? What kinds of environmental challenges (e.g., disease) can best be adapted to through natural selection and thus require maintenance of available genetic diversity? Under what circumstances would populations benefit from introduction of additional genetic diversity? If at-risk populations are to be augmented by importing bears from viable populations, how important is it to assure that the transplants come from similar habitats and/or have similar genotypes? To maintain and enhance adaptedness to current environmental conditions, bear management should aim to preserve both existing genetic diversity of bears, and their stores of knowledge -- assuring that this knowledge is passed from bear to bear, for instance from mother to offspring. For example, highly experienced older females might be given special protection. To maintain adaptability to future conditions, .... genetic, learning .... 7) Inbreeding depression.--If, as expected from theory and initial analyses of tissues, genetic diversity for bears is low, how did this arise and what does it imply about dangers from inbreeding? What has been the role of the founder effect -- whole populations being founded by individuals with little genetic diversity among them? Are mutation rates particularly low? Has natural selection been particularly intense? How much of the diversity represents genetic load; that is, how much range in adaptedness is there between alternative genotypes? Inbreeding can impair current vigor of individuals and populations by increasing phenotypic expression of delterious alleles, for instance alleles that are dominated by beneficial alleles in heterozygotes and expressed only in homozygotes. Inbreeding can also impair adaptability and population viability if it eliminates alternative alleles that may prove superior under future conditions -- a possibility, but not a strong probability. How much would low genetic load and high phenotypic plasticity reduce need for genetic diversity? 8) Ecosystem responses to impacts on bears: More thoroughly determine the roles played by bears in each ecosystem (e.g., predation on ungulate fawns, competition with other carnivores for carrion and occasionally prey, dispersal of fruit seeds, vector for Trichinosis, etc.). Assess how impacts on bears (e.g., declining abundance) would affect each major kind of ecosystem they inhabit; are bears a "keystone" in any ecosystem? 9) Literature review, analysis, and synthesis.--Periodically collate and review existing information on each species of bear (e.g., update the Grizzly Bear Compendium, LeFranc et al. 1987). Then analyze, synthesize, and interpret the information. We have only begun to extract information from existing data, much less to interpret it in terms of general knowledge in biology and ecology. Use this foundation of information to guide management, to identify major gaps in knowledge required for successful bear conservation and management, and to guide research. 10) Standardize data collection and analysis so that all sources can be integrated and compared. Re-evaluate existing data (e.g., for home range size) using standard methods of analysis (including various alternative methods where available) so that similarities and differences between times, populations and species can be verified and interpreted, and effects of method choice evaluated. 11) Accelerate dissemination of information relevant to bear conservation: Assure that new information on bears is available within a few months of production. Keyword index each paper and include the citation in international databases. Currently, most scientific papers on bears are published in the proceedings of the international bear conferences, held triennially; publication is delayed 1-3 years after each conference. Encourage the International Bear Association to reschedule the conferences to meet biennially in North America and Eurasia, but meeting on each continent in alternate years. At each conference, devote some sessions to (a) review papers, (b) integration of information on bears with general scientific knowledge, and (c) conservation/ management strategies. Conferences should serve not only to report recent achievements, but to create new ones on site through "think tank" dialogue sessions addressing critical issues. Foster creation of a new (electronic?) journal and/or bulletin board to publish conference proceedings and other critical bear information (scientific, economic, legal, cultural, etc.) immediately after the information is produced and peer reviewed. Activism Legislation Declining bear numbers led to Threatened status in 1975 for the grizzly bear (Ursus arctos) south of Canada, and recently for some black bear (U. americanus) populations in the southeastern USA. In America, the polar bear is protected by the Marine Mammal Act. These, and comparable measures in other nations have slowed but not reversed declines in bear population size and viability (Servheen 1990). Bear conservation could be further enhanced by modifying existing laws, regulations and policies, and by adding new ones. 1) Promote upgrading and reauthorization of the Endangered Species Act. 2) Foster peer review of management plans for at-risk populations and species. (which peers?) If necessary, mandate this legislatively. 3) Promote structuring of management plans for at-risk populations so that they adequately address legal, then scientific requirements and constraints before addressing political and economic constraints. 4) Foster ecosystem level management of at-risk populations and species. 5) Where necessary to conserve at-risk species and subspecies, promote new laws and/or stiffer penalties to better control (a) sale of bear parts, (b) private ownership of live bears, and (c) transport of bears or bear parts -- with the goal of assuring that they are not traded illegally or destructively. 6) Promote compliance with all laws affording protection to bears by all levels of government -- municipal, county, State, Tribal, and Federal (including USFWS, USFS, FERC, BIA, etc.). 7) Foster development of State and Tribal laws providing protection for bears at least as stringent as Federal laws. 8) Promote adequate consideration of conservation issues in all relevant statutes, regulations, and government policies (e.g., NAFTA). Foster needed refinements of those legal statements. 9) Promote impact review for all human activities -- including national foreign policy, domestic economic policy and private property rights -- which could significantly impact bears. 10) Definitions of key terms (e.g., "habitat", "impact") in government statutes, regulations, and policies should be explicit so as to prevent their redefinition to dilute effectiveness of legislation affording protection to bears. As far as is feasible, definitions should be not only scientifically meaningful, but measurable and precise. Statutes and regulations should be logically self-consistent and specific enough that it is clear what scientific data would suffice to answer key questions -- e.g., whether construction of a ski area on public land would "destroy" critical black bear habitat on an adjacent state forest? (Vague laws are "full employment acts" for attorneys). 11) Enhance protection of "whistle blowers" employed by governments, businesses, and schools (e.g., universities). 12) Expert witnesses: Hiring experts to plan and implement management, to say nothing of doing impact assessments, can be extremely expensive -- so expensive that only governments and corporations can afford it. Rather than let this preclude deep involvement by other conservationists in these processes, ways need to be found to utilize government experts and consultants as though they were one's own employees or allies -- for instance by deriving beneficial testimony from them during litigation hearings. It would behoove conservationists to draw anti-impact testimony from witnesses present to testify on behalf of pro-impact employers/clients ("impactor witnesses"). A hearing is an adversarial confrontation whose purpose is to reach one or more decisions (e.g., whether to allow logging in a specified area, under specified conditions). The only evidence admitted is that allegedly critical to reaching these decisions. Attorneys try to orchestrate testimony and other evidence to build irresistible support for their own case, while rebutting the opponent's case. To this end, each party in litigation normally presents witnesses to support its case, then cross examines opposition witnesses to discredit them. Only rarely is cross examination used to derive testimony from an opposition witness to support one's own case. For opposition witnesses are assumed to be advocates for the party they "represent," not objective, disinterested observers. That assumption is often warranted; often, but not always. Many government scientists and resource managers called to testify are chosen not because of their own beliefs or professional oppinions, but because of their official responsibilities and authorities. A government bear biologist called to testify that logging would not seriously impact a bear population may not actually hold that opinion, although highly specific questions put to him/her by the attorney could give that impression. His/her true opinions might emerge only if appropriate questions are posed under cross examination. The same may be true of professional consultants. Some are definitely "hired guns" who collect data specifically to assist in building the client's case. However, others are objective scientists whose conclusions are dictated by the data. When hired by a client, they cannot predict whether data will favor the client or not. So the client may end up being stuck with data harming his/her case and have to make the best of the situation, either by not having the consultants testify, or by trying to restrict their testimony to points favorable to the client's case. Again, proper cross examination can reveal the true story. Rarely in any kind of litigation is cross examination used to "fish" for information from opposition witnesses. Indeed, a classic guideline for cross examination is to "never ask a question whose answer you can't anticipate, and thus to which you are not prepared to respond." The wisdom of that dictum may be beyond challenged when dealing with witnesses who are advocates for their clients. But it is not necessarily valid when dealing with witnesses who are objective or hostile to the case being made by their employer/client. Hence, it is essential to distinguish the witnesses' position before or at the beginning of cross examination and proceed accordingly. Witnesses cannot properly volunteer information harmful to the case of their employer/client. To do so could be deemed as acting in a "conflict of interest," and perhaps make the witness vulnerable to lawsuit or to judgement as being "in contempt of court." However, a witness can freely answer any questions asked under cross examination. The challenge for a cross examiner is to ask questions that range across enough breadth, yet have enough specificity to allow witnesses to provide a fuller understanding of their findings. For example, during cross examination, "impactor" witnesses might be questioned about whether any of their findings contradict any of the conclusions promoted by the "impactor" (e.g., clients, employers) whom they represent. Activists should thus seek to modify rules for expert testimony, as follows. (a) Assure that witnesses have the opportunity to tell the "whole truth" of their findings, not just the portions selected by attorneys under direct or cross examination. (b) Preferably follow the scientific publication precedent of presenting results before presenting their interpretation. (c) Require that at the end of any given session of deposition, direct or cross-examination testimony, each witness is asked whether his/her testimony has told the "truth, the whole truth, and nothing but the truth" about the issues covered by testimony. If their testimony is incomplete, could that incompleteness bias the Court's understanding of the testimony? For example, if asked whether a given index of reproductive rate is increasing, a witness might truthfully respond "yes," yet believe that the index is misleading. Witnesses should be asked about their reservations about data quality, amount, reliability, representativeness, etc. Data limitations are often imposed by bureaucrats who don't want biologists finding out "wrong information". If weaknesses are to be brought out in court, this would put pressure on bureaucrats to provide better support to biologists gathering data. 13) Promote use of expert judges (Masters) to evaluate the factual basis of each case, leaving legal judges to evaluate the legal basis. 14) Foster tighter specification of the specific kinds of information needed to answer critical questions for assessment of impacts on bear populations and habitat. 15) Promote requirements that research results be accompanied by statements of management implications for conservation (not just for sport harvest), so that managers and administrators can not stifle interpretation of such implications by researchers. 16) Periodically review relevant case law and publish summaries accessible to conservationists. 17) Promote actions to minimize further environmental degradation within, adjacent to, or surrounding National Parks and other prime habitat for bears. Try to confine development to areas outside of prime habitat (e.g., away from travel corridors, food concentrations, denning areas and other critical micro-habitats). 18) Promote legislation to place the burden of proof on parties claiming that their activities will not significantly impact Threatened and Endangered species, or that a species is ready for delisting. Enforcement 1) Funding and other support: Increase where necessary, for instance for forensic procedures to identify body parts from endangered bear species and subspecies. 2) Critique EAs, EISs, forest management plans, and other Federal, State, Tribal, county and municipal documents relating to bear conservation or impacts on bears to assure that NEPA, ESA, and other Federal laws are being enforced. Politics 1) ... 2) ... REFERENCES CITED CEA/HEA Brown, G. 1993. The great bear almanac. Lyons & Burford, Publ., NY. 325pp. Craighead, F. C. 1973. Track of the grizzly. Sierra Club Books. Geist, V. 1991. Phantom subspecies: the wood bison Bison bison "athabascae," Rhoads 1897 is not a valid taxon, but an ecotype. Arctic 44:283-300. ______. 1992. Endangered species and the law. Nature 357:274-276. Hall, E. R. 1984. Geographic variation among brown and grizzly bears (Ursus arctos) in North America. U. Kansas, Mus. Nat. Hist., Spec. Publ. 13:1-14. LeFranc et al. 1987. Grizzly Bear Compendium. Goldman, D., P. R. Giri, and S. J. O'Brien. 19__. Molecular genetic-distance estimates among the URSIDAE as indicated by one- and two-dimensional protein electrophoresis. EvolutioN 42:282-295. Mattson, D. 1991. ______. and J. J. Craighead. 1994. The Yellowstone grizzly bear recovery program: information and uncertainty in endangered species management. in T. W. Clark, R. P. Reading, and A. Clarke, eds. Implementing endangered species policy: reviewing the experience and learning the lessons. Island Press, Washington D.C. ______. and R. R. Knight. 1991. Application of cumulative effects analysis to the Yellowstone grizzly bear population. Interagency Grizzly Bear Study Team Report: 1991C. Merriam, C. H. 1918. Review of the grizzly and big brown bears of North America (genus Ursus) with description of a new genus, Vetularctos. North Amer. Fauna, 41:1-136. ______. 1919. Criteria for the recognition of species and genera. J. Mamm. 1:6-9. O'Brien, S. J., and E. Mayr. 1991. Bureaucratic Mischief: recognizing species and subspecies. Science 251:1187-1188. Rogers, L. L., 1983. ______. 1987. Schaller, G. B. 1993. The last panda. U. Chicago Press, 291pp. Servheen, C. 1991. Shaffer, M. 1980. ______. 1992. Keeping the grizzly bear in the American West: A strategy for real recovery. Wilderness Society Spec. Report. 17pp. Shields, G. F., and T. D. Kocher. 1991. Phylogenetic relationships of North American ursids based on analysis of mitochondrial DNA. Evolution 45:218-221. Stringham, S. F. 1980, 1983, 1985, 1986, 1991a,b, 1992, 1994 U.S.F.W.S. 1993. Wilson, E. O. 1975. Badger 2M Table 1. Hypothetical relationships between road conditions and breaking distances -- used to illustrate by analogy the basic concepts of adaptive management. _____________________________________________________________________ BREAKING DISTANCE Winter tires Winter tires Summer tires ROAD CONDITIONS with chains w/o chains w/o chains _____________________________________________________________________ Dry, clean 25 50 100 Wet, icy 75 150 300 Uncertain 50-75 100-150 200-300 _____________________________________________________________________ A common method of minimizing conflicts is to produce impact studies that spuriously find "no significant impact." "Seek not and ye shall find not." All too often, the only management options addressed explicitly in strategic plans are those deemed politically correct as well as legally compliant. References to scientific literature are used more to rationalize political correctness than to shape a plan itself. When politics are put ahead of legality and science, like carts before horses, management efforts don't carry us to successful bear conservation (cf. Mattson and Craighead 1994). Particularly controversial, at present, are (a) the U. S. Fish & Wildlife Service (USFWS 1993) Grizzly Bear Recovery Plan (GBRP) and (b) the U. S. Forest Service (USFS) Cumulative Effects Analysis and Model (CEA/CEM), with their underlying Habitat Effectiveness Indices (HEI) (REF). There is no evidence that either the GBRP or the CEA/CEM substantially restricts impacts on bears (cf. Mattson 1991); rather, each seems more effective at providing sophisticated rationales for ignoring impact. Such faulty approaches should not be propagated to management of other species within the USA, much less abroad. Management is aided even by crude quantification, such as binary classifications of habitat quality or food supply (e.g., Until detailed plots of this sort are available for changes in each major environmental feature, one can make only educated guesses about how severely environmental changes would impact demographics or habitat effectiveness. habitat degradation can lower reproductive rate and population vigor for bears, and can displace bears to other habitat. Some human impacts can directly kill bears. However, there aas been little detailed description or quantification of the causal linkages. There are many kinds of habitat degradation and many kinds of human impact that can reduce habitat effectiveness or population vigor. We have little quantitative information on how much habitat effectiveness or population vigor are reduced by any specified range of impacts from any single source. For example, how much is population vigor reduced by road densities ranging from 0.1 to 10 miles of road per square miles of landscape. Further insight can be gained by considering a graphic analogy: Suppose that one were assessing potential injury to animals by ultraviolet radiation in circumpolar locations where the atmosphereic ozone layer is thinning. Initially, one might consider indexing injury buy the amount of cell damage. However, it would quickly be apparent that injuring several square centimeters of tissue on the torso or limbs would have far less impact than comparable damage to the surface of the eyes.