Ninth International Conference on Bear Research and Management Monograph Series No.3 (1994)

Density-Dependent Population Regulation of Black, Brown, and Polar Bears

Edited by Mitchell Taylor

with contributions from David L. Garshelis on black bears Bruce McLellan on brown bears Andrew Derocher and Mitchell Taylor on polar bears

An invited paper presented at the Ninth International Conference on Bear Research and Management held at Missoula, Montana, USA February 23-28, 1992

International Association for Bear Research and Management

1994

\ Suggested citations:

Taylor, M., ed. 1994. Density-dependent population regulation in black, brown, and polar bears. lnt. Conf. Bear Res. and Manage. Monogr. Series No.3. 43pp.

Garshelis, D.L. 1994. Density-dependent population regulation of black bears. Pages 3-14 in M. Taylor, ed. Density-dependent population regulation in black, brown, and polar bears. lnt. Conf. Bear Res. and Manage. Monogr. Series No.3. 43pp.

McLellan, B. 1994. Density-dependent population regulation of brown bears. Pages 15-24 in M. Taylor, ed. Density-dependent population regulation in black, brown, and polar bears. lnt. Conf. Bear Res. and Manage. Monogr. Series NO.3. 43pp.

Derocher A.E., and M.K. Taylor. 1994. Density-dependent population regulation of polar bears. Pages 25-30 in M. Taylor, ed. Density-dependent population regulation in black, brown, and polar bears. lnt. Conf. Bear Res. and Manage. Monogr. Series No.3. 43pp.

1994 International Association for Bear Research and Management

Prepared for publication by National Park Service, Yellowstone National Park, Wyoming Printed by Port City Press, Washington, D.C.

ISBN 0-944740-05-7

Available from Michael R. Pelton, Department of Forestry, Wildlife and Fisheries, The University of Tennessee, P.O. Box 1071, Knoxville, TN 37901-1071, USA. Proceedings of this and previous conferences also available. Please write for price list.

\ Density-Dependent Population Regulation of Black, Brown, and Polar Bears

INTRODUCTION .

MANAGEMENT RECOMMENDATIONS 2

DENSITY-DEPENDENT POPULATION REGULATION OF BLACK BEARS...... 3 Documentation of Intraspecific Killing ...... 3 Case Studies With Reported Density Effects ...... 4 Cold Lake, Alberta...... 4 Long Island, Washington , 6 Stockton Island, Wisconsin ...... 6 Interior Highlands of Arkansas ...... 7 Other Evidence For Social Constraints On Population Growth ...... 9 Social Regulation of Subadult Survival ...... 9 Social Regulation of Reproduction ...... 11 Comparisons Among Populations ...... 11 Summary 13

DENSITY-DEPENDENT POPULATION REGULATION OF BROWN BEARS 15 Documentation of Intraspecific Killing 15 Case Studies With Reported Density Effects 16 Yellowstone: 1959-1970 16 Susitna River, Alaska 18 McNeil River, Alaska ...... 19 Glacier National Park, Montana 19 Comparisons Among Populations ...... 20 Summary 23

DENSITY-DEPENDENT POPULATION REGULATION OF POLAR BEARS 25 Case Studies With Possible Density Effects 25 Western Hudson Bay (Churchill, Manitoba) 25 Eastern Beaufort Sea ...... 27 Svalbard 27 Comparisons Among Populations ...... 27 Summary 29

LITERATURE CITED 31

APPENDIX I: DENSITY-DEPENDENT POPULATION REGULATION AND HARVEST POLICIES FOR BEARS 38 Preface

The 9th International Conference on Bear Research and Management was held in Missoula, Montana, February 23-28, 1992. Additional sessions were held in Grenoble, France, October 19-22, 1992. The conference theme was "Management of Bears in a Time of Rapid Change." The cosponsors of the Missoula portion of the conference, whose proceedings are represented here, were the U. S. Department of Agriculture Forest Service, Region One Intermountain Research Station; U.S. Department of Interior National Park Service, Yellowstone National Park; Montana Department of Fish, Wildlife and Parks; U.S. Department of Interior Fish and Wildlife Service; The Wildlife Society; the Interagency Grizzly Bear Committee; and the U.S. Department of Interior Bureau of Land Management. The conference featured 8 sessions of paper presentations plus a poster session, and was attended by more than 400 people. Besides the formal presentations, there were a number of informal special talks and panels, as well as exhibits from bear-management-related industry. Media attention continued throughout the conference, and members of the public attended. Many authors submitted their papers for publication in the conference proceedings. All papers were subjected to thorough peer review; only those successfully passing through this process were published. This monograph is the result of an invited paper presented at the conference. All other published papers are in the conference proceedings, the ninth in a series sponsored by the International Association for Bear Research and Management.

James J. Claar Paul SchuUery

• DENSITY-DEPENDENT POPULATION REGULATION OF BLACK, BROWN, AND POLAR BEARS

MITCHELL TAYLOR, Department of Renewable Resources, 600, 5102 50 Avenue, Yellowknife, Northwest Territories X 1A 358

Abstract: Although all populations are ultimately regulated by density-dependent processes, the range of population densities where density affects vital rates and the mechanisms by which density influences population dynamics have not been demonstrated for any bear population. The per capita rates of birth and death that determine the growth rate and sustainable yield rate of bear populations are panly dependent on the population number. Understanding density effects in bears is complicated by multiple year reproduction schedules, low reproduction potential, physiological and behavioral plasticity, long generation time, large home range, low population densities, and high research costs. Most if not all populations of bears have been reduced from maximum density (carrying capacity) by human induced mortality (i.e., harvest, defense, poaching, or incidental kills). Plausible hypotheses have been advanced regarding the mechanism of population regulation for bears, however re-examination of these studies suggests that alternate explanations are also supported by the available data. Evidence for any general or specific form of density effects for any population is inconclusive. Given this uncertainty, and the likelihood that maximum sustainable yield will occur close to carrying capacity, we recommend that managers assume that no increases in reproduction and no decreases in rates of natural mortality will result from reductions in population numbers, at least until such time that density-dependent mechanisms of population regulation in bears have been documented.

Int. Con!. Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

The goal of our paper is to review the information on to be reduced by density effects because these population regulation of bears and provide parameters affect population-growth rate (evolutionary recommendations on how harvest levels should change fitness) less than adult survival rate. The parameters with changes in population density. This is a most likely to be affected by density are also those most management perspective. The management goal is to likely to be affected by environmental variation. provide use of bear populations within conservation Subadult survival also may be particularly vulnerable limits. The issue is the degree to which the per capita to density effects that are nutritional rather than social. rate of sustainable harvest increases when bear A weaned bear is termed a subadult during the years populations are reduced by hunting. required to reach reproductive maturity. Presumably This issue is an old one in ecology and wildlife the delay in reproductive maturity is to provide the management (Appendix I). All harvested populations individual the time to become more effective at foraging have been reduced to a number that is less than the and to attain a larger body size. A reduction in carrying capacity. However, the degree to which survival from social factors could be caused by additional harvesting will be compensated for by exclusion of subadult males from the home ranges of increased vital rates, and the target population for larger males. However, population growth rate would maximum sustained yield, depends on the mechanism of be impacted less when density effects are mainly density effects (i.e., the shape of the density effects confined to subadult males because the reproductive curve, see Appendix I). potential is not reduced. For bears, there is an array of life-history parameters There are other attributes of bears that hamper that may be affected by density, including age at first studies of density effects. Bears are generalists, reproduction, litter size, fraction of available females exhibiting a wide variety of ecological strategies in that mate and produce cubs, cub survival, yearling different and sometimes highly variable environments survival, subadult survival, and adult survival. (Stirling and Derocher 1990). They have the ability to Additionally, changes in any of these life-history alter their behavior, foraging strategy, life-history parameters may be related to the density of some strategy, and physiology to accommodate their specific sex and age group (e.g., adult males) rather ecological circumstances. Even within a given than total density. Moreover, the manner in which population, researchers are often struck by density affects life-history parameters may differ individualities in behavior. Few populations of North between populations. American bears are free from human-induced mortality The life-history strategy of bears emphasizes long life (Servheen 1990), and thus exist at densities less than to reduce the impact of variability in recruitment rates carrying capacity. If the major effects of density on and cub survival on population growth rate (Taylor et life-history parameters and population-growth rate occur al. 1987a). Cub production and cub survival (both near carrying capacity, then population densities of cubs and yearlings for brown [Ursus arctos] and polar human-impacted populations may be sufficiently low [Ursus maritimus] bears) are the most likely parameters that changes in vital rates and population growth rate 2 Int. Con! Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp. may stem from stochastic (e.g., environmental) factors MANAGEMENT RECOMMENDATIONS rather than density effects. The intrinsically low rate of Bear populations are difficult and expensive to study. increase of most bear populations makes it unlikely that Interest in bears is high and experimental density will change measurably over the time course of manipulations, especially reductions, may not be most studies (i.e., < 10 yr). regarded as ethically acceptable. Virtually all black and Before reviewing the available information, it is brown bear populations and most polar bear populations worth describing some requirements for a successful are impacted by humans, and may not exist at densities study on density effects in bears. There are 2 sufficient to demonstrate density effects. Investigation approaches that could be considered: mechanistic (the of density-dependent population regulation of harvested biology of how density affects rates of birth and death) populations would require a prohibition of harvest and and empirical (correlating changes in population growth a long-term study to allow the population to increase rate with density). Both types of study would have to sufficiently for density effects to become apparent. The span the range of densities where vital rates were reproduction schedule of bears is flexible to reduced. A mechanistic study would involve annual accommodate environmental variability, and bear estimates of population density and rates of birth and behavior is very adaptive to new and novel conditions. death for the full complement of bear population Detailed studies of food availability also would be parameters. The relationship between population necessary to rule out density-independent fluctuations in density and vital rates would be determined (e.g., Fig. the environment. These constraints have thus far 11 in Appendix I) for a given population. The carrying prevented an unambiguous demonstration of the capacity could be estimated by extrapolating the mechanism of density-dependent regulation of any observed relationships to a density where population population of bears. growth rate is equal to one. The empirical approach In the 3 species-specific reports that follow, there is would be restricted to estimates of population densities no evidence that fluctuations in survival and recruitment (and subsequently population growth rate) through a rates were caused by density of bears (or density of range of population-growth rates sufficient to document male bears). Nevertheless, density-related reductions in the shape of the density-effects curve. Both approaches cub survival, especially cubs in their first year, was the would allow a manager to determine the population most suspected mechanism of density effects. This density at which the highest sustained yield would could have been caused directly by intraspecific occur. With these constraints in mind, we consider the aggression or indirectly by exclusion from prime available information on density effects for black (Ursus habitats. Forced dispersal of subadult males was also americanus), brown, and polar bears. suggested; however, to regulate the population an effect We thank S. Miller for suggesting this topic to us on females would also be required. The existing data and for his comments on an earlier draft. J. Troje were insufficient to distinguish between social or prepared the figures, L. Self contributed to word nutritional effects for any of the studies considered with processing the manuscript, and S. Broadbent did final the exception of 1 black bear study (Clark 1991). editing and layout. We acknowledge R.K. Anderson, Given these uncertainties, and the theoretical J.D. Clark, K.D. Elowe, S.P. French, D.P. Fuller, expectation that density-dependent compensatory G.B. Kolenosky, A.L. LeCount, D.L.D. Martinez, increases in vital rates will occur mainly at densities C.R. McLaughlin, P.K. McLean, M.R. Pelton, D. close to carrying capacity (Appendix I), we join Miller Robertson, L.L. Rogers, 1. Schoen, C. Shank, 1. (l990b) in recommending that managers of harvested Stirling, S. Stringham, A. Sutherland, J. Tees, D.R. populations not base sustainable yield projections on Trauba, L.G. Visser, A. Welch, J.B. Wooding, and 2 anticipated increases in vital rates that might occur from anonymous reviewers for unpublished data and/or their harvest reductions in population density. A fuller comments on all or portions of earlier drafts. understanding of these issues will require long-term, This work was supported by the Northwest intensive studies over the range of densities where Territories Department of Renewable Resources, regulatory effects are operative. Such studies must not Minnesota Department of Natural Resources, British only estimate vital rates and population densities, but Columbia Ministry of Forests, and Canadian Wildlife also measure (or index) habitat quality (i.e., food Service. Our particular thanks to the unnamed field supply) and determine the mechanism by which density biologists, technicians, and students who collected and causes changes in vital rates. Until data from such tabulated the data that were used in these analyses. studies are available, our understanding of population regulation of bears will remain largely conjecture. DENSITY EFFECTS ON BEARS • Taylor et al. 3

DENSITY-DEPENDENT POPULATION REGULATION OF BLACK BEARS

DAVID L. GARSHELlS, Minnesota Department of Natural Resources, 1201 East Highway 2, Grand Rapids, MN 55744

The form, significance, and even existence of and consequent effects on survival and reproduction. In density-dependent processes in animal population the first of these I examine case studies with reported dynamics have been debated among ecologists for density effects. Next, I discuss the collective evidence decades (Krebs 1978; Murray 1982; Strong 1986,1987; from a host of other studies regarding the social DenBoer 1987; Hassell and Sabelis 1987; Lomnicki regulation of survival and reproduction. 1987; Appendix I), and are cause for controversy and Finally, in the fourth section under the heading of uncertainty in black bear management programs "comparisons among populations," I probed for (California Department of Fish and Game 1991, Gill relationships between population density or sex-age and Beck 1991). It seems logical to assume that composition and corresponding rates of reproduction increasing bear densities will eventually retard fecundity and survival in a data set assembled from black bear and/or survival, so conversely, these vital rates should studies conducted throughout North America. improve if density is reduced through harvest. Thus, harvest mortality might be expected to be at least partially compensated for by enhanced reproduction or DOCUMENTATION diminished natural mortality. This scenario not only OF INTRASPECIFIC KILLING provides an intuitive explanation for populations Intraspecific killing is the most direct potential achieving relative stability (Royama 1977), but as mechanism of population regulation in black bears, but Miller (1990b) observed, compensatory mortality has a considering the multitude of telemetry studies that have seductive appeal to management agencies because it been done on this species, the number of observations embodies inherent safeguards against over-harvest. of intraspecific killing have been surprisingly few. Nevertheless, density-dependent effects are difficult to Several studies documented cases of black bears killing detect, even in populations where animal abundance can other (typically young) bears in traps (Erickson 1957; be accurately enumerated (Murdoch 1970, Vickery and Jonkel and Cowan 1971; Payne 1978; Kemp 1976; Nudds 1984, Hassell 1986, Pollard et a!. 1987, Kautz Lindzey et a!. 1986; Rogers 1987a; Clark 1991; 1990), and obviously more so for species like black McLean 1991; D.M. Graber, Nat. Park Serv., pers. bears that not only elude accurate numerical estimation, commun.; A.L. LeCount, Ariz. Game and Fish, pers. but also defy rigorous monitoring of population trend commun.; G.B. Kolenosky, Ont. Minist. Nat. Resour., (Miller 1990d; Garshelis 1991, 1993). pers. commun.; C.R. McLaughlin, Me. Dep. Inland This review evaluates present evidence for density­ Fish and Wild!., pers. commun.), but these incidents dependent responses in black bear populations. The may be only remotely related to what occurs in the intent here is not to determine whether, when, or how wild. From reports in the early literature (prior to density-dependent processes are manifested, but rather 1970), Rogers (l987a) tabulated only 3 cases of wild to compile and assess the available information, and black bears killing cubs and 1 case of a bear killing a thereby clarify our current level of understanding of this mother with 2 cubs in a den. During the past 2 issue. decades, an average of about 2 bears per year have This review is divided into 4 main sections. I begin been reported killed by other bears among the various with a tabulation of cases of intraspecific killing. studies conducted across North America. The most Corresponding density estimates were not available for common victims (17/39 or 44 %) were adult females most of these cases and the total number of cases is (Table 1), which is likely related in part to the emphasis small, so relationships between killing and bear density on reproduction in black bear telemetry studies, and could not be examined. However, the small number of hence the preponderance of adult females among radio­ cases itself indicates that intraspecific killing is likely an marked animals. unimportant mechanism of population regulation, at It is often presumed that most killing is done by adult least at the densities that black bears currently exist. males, although it is not intuitive that adult males The next 2 sections discuss alternate potential should kill potential mates. It may be that some mechanisms of population regulation, including social females were killed trying to protect their cubs, as there constraints on immigration, emigration and habitat use, are clear advantages for males to kill unrelated cubs in 4 Int. Conf Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

Table 1. Intraspecific killings of black bears (excluding cubs) females, I with cubs (G.B. Kolenosky, Ont. Minist. documented among bear studies across North America, 1975­ Nat. Resour., pers. commun.; L.L. Rogers, U.S. Dep. 92. Agric. For. Serv., pers. commun.). Adult females may Sex/age Free- kill other females in overlapping home ranges as a of victim ranging In den Total Source(s) result of resource competition, or where there is little Adult female with cubs 6 4 10 a,e,i,j,k,l,m,n,p overlap among female home ranges, females may kill adjacent unrelated territory holders to enable their own Adult female with 2 0 2 p female progeny to establish a territory. yearlings Although many studies reported significant rates of Barren adult female 3 2 5 d,f,h,i,o cub mortality (based on their absence from the mother's Yearling with mother 0 den a year after birth), few were able to document (mother not killed) causes of death. Thus, although intraspecific killing of Subadult male (1-3 yrs) 9 10 b,g,i,j,k,n,o non-cub bears is apparently rare, cub-killing may be Subadult female 11 0 11 b,c,l,m,q significant but largely unidentified because cubs are normally not radiomarked. LeCount (1987a) found that Total 32 7 39 adult males were the primary predator and principal cause of death among radio-collared cubs in Arizona. a Alt and Gruttadauria (1984). b Anderson and Trauba (1991) and D.R. Trauba (Vniv. Wis., Similarly, black bear predation was the second-most Stevens Point, pers. commun.) (1990-91 observations). common cause of death for collared orphan cubs in c Beecham (1980b) (1975 observation). Ontario (Strathearn et al. 1986). C. R. McLaughlin d D.L. Doan Martinez (Texas A&M Vniv., Kingsville, pers. (Me. Dep. Inland Fish and Wildl., pers. commun.) also commun.) (January 1992 observation; cub in utero). reported a case of a collared orphaned cub being killed e K.D. Elowe (Me. Dep. Inland Fish and Wild!., pers commun.) (1983 observation). by another bear in Maine. Conversely, except for 2 f S.P. French and M.G. French (Yellowstone Grizzly Foundation, cubs eaten by their mother, Elowe and Dodge (1989) pers. commun.) (1988 observation). observed no cases (in 19 ascertained causes of death) of g D.L. Garshelis et al. (Minn. Dep. Nat. Resour., unpub!. data) intraspecific killing of radio-collared cubs in (1989-92 observations). Massachusetts. Lindzey and Meslow (1977a) and h Hellgren and Vaughan (1989) (1986 observation). G.B. Kolenosky (Ont. Minist. Nat. Resour., pers. commun.) leCount (1982, 1990) suggested that comparatively (1990-91 observations). high cub mortality in years when relatively few females LeCount (1982, 1987, 1990, Ariz. Game and Fish, pers. were accompanied by cubs was evidence for cub commun.) (1976-1989 observations). predation by barren adult females. However sample k Lindzey et al. (1986) (1980-1982 observations). I C.R. McLaughlin (Me. Dep. Inland Fish and Wild!., pers. sizes were small in these studies and no other studies commun.) (1983-89 observations). have presented corroborative results. m L.L. Rogers (V .S. Dep. Agric. For. Serv., pers. commun.) (1987­ 1989 observations). CASE STUDIES WITH n Schwanz and Franzmann (1991) (1983 observation). REPORTED DENSITY EFFECTS a Tietje et a!. (1986) (1976 observations). P L.G. Visser (Mich. Dep. Nat. Resour., pers. commun.) (1988 observations: cause of death of adult female with yearlings less Cold Lake, Alberta certain than that of adult female with cubs). To test the hypothesis that interactions between adult q Wooding and Hardisky (1994) (1986 observation). males and younger bears contribute to the regulation of black bear populations, Kemp (1972, 1976) initiated the hope of later mating with the mother. In 7 cases what has now become a legendary experimental the killer was known from telemetry or direct visual manipulation of a previously unexploited population in observation (Tietje et al. 1986; Schwartz and east-central Alberta. After observing relatively stable Franzmann 1991; S.P. French, Yellowstone Grizzly population estimates for the first 4 years, Kemp Foundation, pers. commun.; Garshelis, Minn. Dep. removed 23 adult (and 2 large subadult) males from the Nat. Resour., unpubl. data), or strongly suspected study area over a 2-year period (1971-72). He based on other sign (Tietje et al. 1986, Hellgren and estimated that the population increased from a mean of 2 2 Vaughan 1989, Wooding and Hardisky 1994) to be an 80 bears in the 218-km study area (37 bears/lOO km ) adult male; in 5 of these cases the victim was a female, during the pre-removal period to 117 in 1972 and 175 3 of which were adults. In 3 other cases adult females, in 1973. Young and Ruff (1982) continued the study 2 of them with cubs, were the known killers of other and noted a decline to 136 and 137 bears in 1974 and

• DENSITY EFFECTS ON BEARS • Taylor et al. 5

1975, respectively, and 2 years later the population had caught after the adult removal may be an indication that apparently returned to the pre-removal level. the perceived population increase was not just a Concurrent with the population increase was a manifestation of changing capture probabilities. substantial increase in the percentage of subadults (22 % However, even if we tentatively accept the hypothesis pre-removal versus 48% post-removal). Additionally, that more subadults entered the study area after the among the subadults, the sex ratio shifted dramatically removal of the adults, it is impossible to tell whether toward males. These observations prompted the these animals settled there. Unlike more recent black conclusion that adult males had been regulating bear studies, bears captured at Cold Lake were not population density via subadult recruitment. Removal radiomarked. If they had been, their departure (either of a large number of males seemed to cause either an temporary or permanent) from the study area could influx of young males from outside the study area have been determined telemetrically, and the population and/or reduced loss (emigration and mortality) of young estimates adjusted accordingly (e.g., Garshelis 1992). males born on the study area. Young and Ruff (1982) radiocollared bears beginning in Results of this study have been embraced as virtual 1974, but there is no indication that they utilized dogma, partly because the experimental design appeared telemetry data to determine the number of marks so elegant (and so rare among bear studies), and partly available for capture. By using the Lincoln-Petersen because the conclusions seemed so logical in the context estimator, the Cold Lake investigators tacitly assumed of what was and is now known about bears. However, that none of the animals marked in the previous despite their intuitive appeal, these conclusions may be trapping period left the area; if in actuality, many of the unwarranted. newly captured subadults merely passed through the Kemp (1972, 1976) and Young and Ruff (1982) area, they would nonetheless have been tabulated in the relied principally on the Lincoln-Petersen mark­ marked population (M) available for capture during the recapture estimator to derive population estimates. This next marking period, thereby inflating the population approach considers all previously-marked animals to be estimate for that period (N = (Men)lm). That is, by available for capture; that is, it assumes the population using an estimator that assumes closure on a population is closed (marked animals do not leave). Furthermore, that has become less closed due to the disruption of the it assumes that all animals have an equal probability of residents, estimates of population size would capture. Both of these assumptions appear to have been correspondingly increase. In fact, the number of bears violated in this study, resulting not only in severely present on the area at any moment in time might not biased population estimates, but possibly even an have changed at all, and the number of permanent incorrect interpretation of population trend. residents could just as well have decreased. In examining the data presented in these reports, it is Kemp (1976) reported that the recapture rate of evident that adults were recaptured more frequently subadults increased after the removal, and suggeste'N than subadults, either indicating higher emigration for that this was evidence that these animals did in fact subadults or a more negative response to traps. In settle in the area. I reexamined these data and found either case, probability of capture was obviously that the recapture rate increased only during the first different for adults and subadults, so it was not valid to year of the removals (recapture in 1972 of bears caught combine these 2 age groups in application of the in 1971), but fell sharply the following year to the pre­ Lincoln-Petersen estimate. By doing so, any change in removal level. Furthermore, the temporary increase in the relative proportion of adults to subadults in the recapture rate also was observed for resident adult population would necessarily produce a change in the animals (i.e., adults caught and marked prior to 1971, population estimate. Thus, by removing a large and recaptured in both 1971 and 1972). Thus, it number of resident adult males, the ratio of "m" appears that the probability of capture was higher for marked animals captured to "M" marked animals in the all bears in 1972 (possibly due to food conditions or population would likely decrease, commensurate with other environmental factors), which resulted in both the reduced probability of capturing young males. As more total captures and more recaptures that year; these a direct consequence, even with no change in actual data are thus insufficient to infer a change in population size, the estimate of the population (N), emigration, immigration, and settling rates as a result derived by multiplying the inverse of the capture of the removal. probability (Mlm) times the number of animals captured Two other population estimators were used in the (n), would increase. Cold Lake study, which were averaged with 3-6 annual The substantially increased number of subadults Lincoln-Petersen estimates. One was based on 6 Int. Con! Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

estimating the number of bears never captured by fitting resulling in a population decline. a Poisson distribution to the observed frequency of Several causes for the population decline were captures. Roff (1973) has shown that this method is identified, including diminished cub production, reduced also quite sensitive to changes in probability of capture, cub survival, increased dispersal, and increased and moreover, although this method is commonly used mortality. While the population was increasing in the (and was used in the Cold Lake study) to test for equal early years of the study, only males dispersed catchability, it is rarely able to detect capture (permanently) to the mainland (a minimum distance of heterogeneity. Furthermore, like the Lincoln-Petersen only 300 m). However, coincident with reduced berry estimator, this method assumes population closure. production, dispersal of young females from the island The third method used to estimate population size at became more common. Additionally, long-time Cold Lake was based on what was termed "capture resident adults of both sexes began to include parts of efficiency," patterned after Meslow and Keith (1968). the mainland within their home ranges, and 2 adult This was the only method of the 3 that does not assume females eventually moved off the island; adult dispersal, closure. It is a limited form of the Jolly-Seber especially by females, is a rare event in most black bear equation, which is currently the most accepted model populations, as indicated by the absence of such for estimation of demographically open populations observations in other studies. Concurrent with this without using telemetry. It is revealing that population apparent increased rate of dispersal were several estimates derived from the capture efficiency procedure incidents of outright killing of bears (4 free-ranging and show no trend related to the experimental removal 2 in snares) by conspecifics, which was previously not (mean of 2 years pre-removal = 80, mean of 2 years observed on the island. By the end of the study, the 2 post-removal = 76). However, since these estimates estimated density of bears had dropped to 1051100 km . were averaged with the other, more biased estimates, they merely dampened the perception of a population Stockton Island, Wisconsin increase. Stockton Island, one of the Apostle Islands in Lake 2 This reinvestigation of the Cold Lake study suggests Superior, is twice the size (41 km ) of Long Island, that the original conclusions were unjustified and quite Washington, and considerably further from the possibly erroneous. A more thorough investigation of mainland (7.8 km) (but only 2.2 km from the nearest the capture data from this study is in order, using an other island). Thus, one might expect that bears would open population model, such as the Jolly-Seber be less prone to disperse from this island, so increased procedure refined by Pollock et al. (1990), which can density could potentially have a more direct effect on account for unequal capture probabilities. reproduction and survival. The first documented bear activity on Stockton Island Long Island, Washington was reported in 1973 (Jones 1987, cited in Anderson Population dynamics of black bears on a 2l-km2 and Trauba 1991), although sign and sightings were island in southwestern Washington were investigated sporadic through the 1970s. An adult female was from 1973 to 1982 (Lindzey and Meslow 1977a, captured on the island in 1982, but this animal was Lindzey et al. 1986). Population size was estimated translocated as a nuisance and died soon afterwards. In from trapping, telemetry, and visual observations; 1984 an adult male and an adult female that were captures and observations of unmarked adults in the together during the breeding season were captured and latter years of the study were used to help correct radiomarked, and later the progeny of this female were estimates for the early years. It was assumed that by radiocollared. A trapping program was initiated in the end of the study. virtually all animals were 1988 at which time there were an estimated 12 bears on accounted for. the island (Anderson 1989). By 1990 the population Population density peaked in the mid-1970s at 176 had increased to at least 20 bears (excluding cubs) (49 2 2 bears (excluding cubs)1100 km , which is the highest bearsllOO km ). Hunting was prohibited and only 1 density yet reported for black bears, excluding death was recorded on the island during 1988-90, a temporary congregations at concentrated food sources yearling that apparently was killed by another bear in like salmon streams or garbage dumps. During these 1990 (D.R. Trauba, Univ. Wis., Stevens Point, pers. peak years natural foods were abundant. However, as commun.). the clearcuts from former logging operations matured, In 1991, 3 other independent yearlings from 2 litters berry production declined by about 50 %, which also were killed by other bears. All of these yearlings apparently reduced the carrying capacity for bears, were lightweight (11-12 kg) when handled in the den I

• DENSITY EFFECTS ON BEARS • Taylor et al. 7 the previous winter, but 1 survived until early July and Clark (1991) hypothesized that the exceptionally low the other 2 through mid-August (when their carcasses cub survival at White Rock was due to cannibalism by were found within 125 m of each other), making it adult males. Bear densities were relatively low in both unlikely that they starved and were subsequently areas (7.5 noncub bears/l00 km2 at White Rock [171100 scavenged. Another exceptionally lightweight (6 kg) km2 in the densest part of this area], 9.0/1002 at Dry yearling that was not radiocollared presumably died of Creek [23/100 km 2 in the densest area]); however, the starvation in 1991, and it appeared that at least 3 cubs high male:female ratio at White Rock increased the from a particularly lightweight 4-cub litter born in 1991 likelihood of male encounters with cubs that were not did not survive that year (Anderson and Trauba 1991). their own, which may have prompted cub-killing to No permanent dispersal was observed, but 2 males enhance breeding opportunities. Six of 8 entire litters occasionally left and subsequently returned to the were lost on White Rock compared to 0 of 8 at Dry island; one of these made regular trips to another island Creek. and once to the mainland where he was killed by a Females losing litters during the summer came into hunter in 1991. oestrus about a month later than the typical breeding Initial evidence indicated that increasing density period, and Clark (1991) suggested that this delay might corresponded with decreased yearling weights and an account for the smaller average litters produced at increased rate of mortality, due to cannibalism, White Rock (possibly because of reduced follicular starvation, and more regular and extended movements development later in the summer). Known first litters off the island, resulting in heightened exposure to among 3- and 4-year-old females were expectedly small hunting. In a companion telemetry study on the and similar on the 2 areas (x = 1.5 and 1.8 cubs for mainland, where bears were hunted, the density was White Rock and Dry Creek respectively; calculated lower, home ranges were larger but with less overlap, from Clark [1991]), whereas females that lost entire yearlings were heavier, and no natural mortalities were litters on White Rock bore unusually small litters (x = observed. However, this work has not as yet continued 1.2 cubs) the following year. for a sufficient period of time and sample sizes are as There was no documented evidence of cannibalism in yet too small to determine (1) whether the increased this study (except for a single scat containing bear hair), natural mortality observed in 1991 represented the so the proffered scenario remains speculative. beginning of a trend, and (2) whether this increase was However, there was no difference between the areas in actually attributable to effects of relatively high density the age structure of females or their ages of first rather than other chance events (e.g., localized food reproduction. The early age of first reproduction in shortage, presence of a particularly predatory bear). both areas (66% at 3 years old) and correspondingly heavy, and similar weights for bears in the 2 study Interior Highlands of Arkansas areas ruled out a nutritional explanation for the Clark (1991) compared the population dynamics of 2 observed differences in cub survival. Blueberry allopatric bear populations in Arkansas, 1 in the Ozark production was higher on Dry Creek, but hard mast and Mountains and 1 in the Ouachita Mountains. Both most other soft mast was more abundant on White populations had been previously extirpated, and then Rock. Cubs that survived their first year were reintroduced during 1959-68. Hunting was allowed exceptionally heavy in both areas (x = 38 kg at 1 year beginning in 1980. Probably as a result of the timing old), compared to other places in North America of the hunt, a high proportion of females (76%) were (Table 2). taken in the White Rock Wildlife Management Area in Some genetic differences between the 2 populations the Ozarks during the 1980s; field data and were evidenced by differences in coat color, but if poor corroborative modelling indicated that this female cub survival on White Rock was genetically-linked, it biased mortality skewed the resident population presumably would have been manifested throughout the significantly toward males (59% of both adults and historic growth of this population; however, modelling subadults were male). The other study area, referred to efforts, starting with an almost known population (from as Dry Creek, had a sex ratio that favored females reintroduction) of bears in the late 1960s indicated (only 39 % male). Both mean litter size (1.41 versus that the initial population-growth rate at White Rock 2.25) and cub survival (31 % versus 90%) were must have been much higher than at present to produce appreciably lower at White Rock, resulting in a stable the numbers of bears currently residing there. An White Rock population versus a growing (\ = 1.26) unidentified disease affecting cubs (but not population in Dry Creek. older bears) at White Rock has not been discounted, but 8 Int. Conf. Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

Table 2. Characteristics* of black bear populations from across North America.

BearsllOOkm' Annual survival(%) Cub litter

Excluding Including Interval Age 1st Yrlng wt Study area cubs cubs Ad M Ad M:F AdM Cub (no.) (yrs) repro (yrs) (kg)

Susitna, Alas.a 7 9 68 2.1 2.7 5.9 11

White Rock, Ark.b 8 9 2.7 1.45 95 31 1.4 3.3 40

Dry Creek, Ark. b 9 12 1.7 0.59 85 90 2.3 3.3 37

Black Mesa, Colo.c 11 14 1.2 0.38 70 56 2.0 2.2 4.7 23

Leonard Canyon, Ariz.d 11 15 3.7 1.05 55 1.8 2.0 4.6 23

Kenai 1947 bum, A1as. e 15 21 3.7 0.68 90 74 2.2 2.2 5.8 18

Northeastern MinnJ 16 21 2.6 0.51 75 2.4 2.3 6.3 18

Bradford, Me.g 19 30 2.9 0.45 61 29 2.4 2.1 4.6 16

Stacyville, Me.g 73 2.6 2.1 4.8 17

Spectacled Pond, Me.g 20 36 4.0 0.42 60 62 2.5 2.0 5.4 13

North-central Minn.h 20 26 2.3 0.48 74 85 2.5 2.1 4.7 21

Kenai 1969 burn, Alas.e 20 27 3.6 0.56 77 91 2.3 2.1 4.6 24

Drummond Is., Mich.i 25 35 5.4 0.59 79 59 2.2 2.0 4.6 20

Western Mass. j 27 39 2.8 0.25 73 53 2.2 2.0 3.7 20

Smoky Mountains NP, Tenn.k 29 7.3 1.06 62 2.6 2.4 4.8 11

White River NWR, Ark.' 36 41 13.5 1.43 95 68 2.3 2.4 4.9

Lowell, Idahom 43 0.97 1.7 5.0

Four Peaks, Ariz.n 36 49 13.7 1.05 58 2.0 2.3 4.7 22

Big Creek, Mont. o 40 45 8.0 0.68 1.7 3.0 7.0

East-central Ont.p 48 57 7.6 0.43 53 2.5 2.1 5.7 16

Stockton Is.. Wis.q 53 64 17.3 0.86 97 69 2.6 2.3 5.0 22

Dismal Swamp NWR, Va., N.C. r 59 21.8 2.50 59 75 2.1 4.2

Council, Id.s 62 77 9.6 0.72 1.9 2.4 4.8

Shenandoah NP, Va.t 86 14.6 0.95 59 70 2.0 2.0 4.0

Long Island, Wash.u 130 162 31.7 0.46 100 27 2.2

Yosemite NP, Cal.v 0.84 2.0 2.5 4.2 29

Yosemite NP, Cal. w 0.23 1.6 4.1

Northeastern Pa.x 84 3.0 2.0 3.2 33

• Mark-recaprure samples were used for most density estimates; adult male (Ad M) densities may have been inflated in many cases due to male­ hiased caprures. Ratios of adult males per female (M:F) also were generally derived from caprure samples, and thus could be similarly male­ biased. Adults were considered to be ~4 years old. Adult male survival was determined from telemetry records. Most male mortality was due to hunting, but male mortality rates were high even in some unhunted populations. All national parks (NP) and national wildlife refuges DENSITY EFFECTS ON BEARS • Taylor et at. 9

(NWR) were unhunted (although males were susceptible to hunting outside), as were White Rock, Dry Creek, Black Mesa, and Long Island during the years indicated; Four Peaks was very lightly hunted « 2 % mortality). Cub survival, litter sizes, and intervals between litters were typically detennined from den checks (except Susitna, Big Creek, and Yosemite) and the inter-litter interval excluded cub production following whole-litter loss. Yearling weights were obtained in the den (13-14 months of age), or shortly thereafter (Susitna and Yosemite), and used as an indicator of food availability. Values for all parameters were either taken directly from the indicated sources or derived from data presented therein. a Miller (1987,1994), Miller et al. (1987). b Clark (1991). c Beck (1991). d LeCount (1987b; Ariz. Game and Fish, pers. comrnun.) (both sides of Mogollon Rim combined). e Schwartz and Franzmann (1991). f Rogers (1987a). g McLaughlin et al. (1994), C.R. McLaughlin (Me. Dep. Inland Fish and Wildl., pers. comrnun.). h D.L. Garshelis (Minn. Dep. Nat. Resour., unpub!. data). L.G. Visser (Mich. Dep. Nat. Resour., pers. comrnun.). j D.P. Fuller (Univ. Mass., pers. comrnun.). k Eiler et al. (1989), McLean (1991), McLean and Pelton (1994; Univ. Tenn, pers. comrnun.). I Smith (1985). m Beecham (1980b; Idaho Fish and Game, pers. comrnun.). n LeCount (1982; 1984; Ariz. Game and Fish, pers. comrnun.). o Jonke1 and Cowan (1971). p Koienosky (1986, 1990), Yodzis and Kolenosky (1986)(1975-80 data only). q D.R. Trauba and R.K. Anderson (Univ. Wis., Stevens Point, pers. comrnun.). r Hellgren and Vaughan (1989). s Beecham (1980a; 1983; Idaho Fish and Game, pers. comrnun.), Reynolds and Beecham (1980). t Carney (1985).

U Lindzey et al. (1986)(1980-82 data only). v Graber (1981) (developed areas). w Keay (1990) (wilderness areas). x Alt (1980, 1981, 1989). seems unlikely. Thus, the heavily male-biased adult home range. Regardless of what instigates males to sex ratio presents the most logically consistent disperse, this behavior could retard population growth explanation for the high cub mortality at White Rock. in a pseudo-density-dependent fashion if (1) dispersal It should be emphasized, however, that cub mortality in entails a heightened risk of mortality, and (2) this risk this area was not related to bear density, so whereas it increases with density. However, increased mortality may be considered a population-composition effect, it of dispersing subadult males would not be sufficient to is not actually an example of density dependence. regulate population size (true density-dependence), unless there were also repercussions for females. Several studies suggested that dispersing bears are OTHER EVIDENCE FOR SOCIAL especially prone to come in contact with humans, and CONSTRAINTS ON POPULATION GROWTH thereby risk being hit by a car, killed as a nuisance, or shot by a hunter (Jonkel and Cowan 1971, Ruff 1982, Social Regulation of Subadult Survival Klenner 1987, Rogers 1987a, Garshelis 1989, Mattson Male-biased dispersal has been well documented in 1990, Schwartz and Franzmann 1992). Currently, there black bear studies (Alt 1978, Hugie 1982, LeCount is no evidence that animals disperse to the periphery of 1982, Beecham 1983, Rogers 1987a,b, Garshelis et al. their range and perish from malnutrition or predation, 1988, Beck 1991, Schwartz and Franzmann 1992). so one could argue that without humans, dispersal does Whereas subadult males typically disperse and not entail increased mortality. However, since humans eventually settle in an area beyond their natal range, are now present within the dispersal range of nearly subadult females tend to settle in or near their mother's every male black bear born in North America, the point 10 Int. Conf Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp. is moot. The really pertinent question is whether dispersal­ related mortality rates correspond positively with A) density. Dominance hierarchies have been observed among male black bears (Herrero 1983, Barber and Lindzey 1986), but it is unclear whether adult males induce juvenile dispersal and/or influence their settling. If subadult dispersal is provoked by adults, then removal of adults could be partly compensated for by • increased retention and lower mortality of subadults (Fig. 1C). Alternatively, subadult males may be B) innately prone to disperse (Rogers 1987a,b), but avoid • areas used by older males when searching for a place to settle. Accordingly, dispersing subadults often settle in habitats where they are more exposed to the risks of 1 human contact, such as dumps (Young and Ruff 1982, Tietje and Ruff 1983, [but see Pelchat and Ruff 1986]), agricultural or residential areas (Rogers 1987a, Mattson C) 1990), areas near roads (Schwartz and Franzmann • • 1992), or areas beyond the confines of sanctuaries like national parks (Beeman and Pelton 1980, Garshelis and Pelton 1981). In this scenario, reduction of adult males would enable subadults to settle in safer places, thereby 1 increasing their survival (Fig. 1B). Finally, it is possible that adult bears do not appreciably affect either the emigration or immigration of subadults. Subadults 0 - Subadult a may naturally disperse, and as an inexorable KEY { consequence of wandering, dispersers may more readily .- Adult a happen upon or be attracted to human-related food sources where they incur an increased risk of mortality. Fig. 1. Mortality of subadult black bears may be heightened In this case, reduced adult density would not improve by their wanderings into human-related mortality sinks (A). If juvenile survival. adult males do not instigate dispersal but preclude dispersing Regrettably, the dispersal patterns of bears are not subadults from settling. then reduced density of adult males may reduce some dispersal-related mortality; however. some understood sufficiently to predict the effects of an dispersing bears will still wander into mortality sinks before altered social structure on subadult mortality. Although finding a safer place to settle (8). If adult males are the Cold Lake study is often touted as the best evidence responsible for prompting dispersal. reduction in adult males that adult males control the settling rates of subadults, should yield significantly fewer dispersal-related mortalities my reanalysis indicated that the results of this study (C). Present evidence seems to support scenario 8 over C. were inconclusive. Furthermore, even if the Cold Lake population did increase after the adult male removal, of exploitation nor yearly population sizes were this study could not differentiate whether the population quantified. Rogers (1987a,b) reported that settling rates increase was due to reduced emigration, increased of young males immigrating onto his study area in immigration, or both. The surrounding area was northeastern Minnesota were higher during a year when unhunted, which certainly hampers the extension of the adult male presence was particularly low, and Garshelis purported findings of this study to the typical case (Minn. Dep. Nat. Resour., unpubl. data) observed a where large unhunted reservoirs of bears are not similar increase in subadult male density in north­ available. Beecham (1983) suggested that because his central Minnesota the summer following a year when study area in Idaho had no nearby bear reservoirs, adult male density was reduced; however, neither study removal by hunters of a "large proportion" of the adult was able to document that the survival of these subadult males did not result in a population increase; immigrants was higher than it would have been had unfortunately, like the Cold Lake study, Beecham's they settled elsewhere. results also were inconclusive because neither the level Even if it could be demonstrated that juvenile male DENSITY EFFECTS ON BEARS • Taylor et at. 11 survival was inversely related to adult male density, females would force some individuals to forage in population consequences of trading adults for subadults nutritionally depauperate habitats. Accordingly, are uncertain. LeCount (1987a) suggested that elevated density could theoretically depress immigrant males are more likely to be infanticidal than reproduction, and conversely reduced density might residents, because residents should not risk killing cubs enhance reproduction. Lindzey et al. (1983) reported that they may have sired. Consequently, if removal of a high reproductive rate for black bears in adult males results in higher densities of newly­ Pennsylvania, which they attributed in part as a maturing subadults, cub recruitment may decline. compensatory response to heavy bear harvests that Alternately, in a hunted population, increased density of reduced density. However, they never documented a subadult males could buffer hunting mortality on change in either bear density or reproduction, and later females, enabling more cubs to be born. These evidence indicated that the exceptionally high potential interrelationships among social factors, sex-age reproductive output was more likely related to plentiful specific survival rates, and population growth are supplies of both natural and human-related foods (Alt evidently too complex to enable accurate predictions of 1989). human-induced perturbations.

Social Regulation of Reproduction COMPARISONS AMONG POPULATIONS Reproduction in black bears is generally thought to A paucity of evidence demonstrating density-related vary in a density-independent manner, due largely to effects on population growth may indicate that such fluctuations in availability of natural food (Rogers effects are not significant at the densities and conditions 1987a, Eiler et al. 1989, Elowe and Dodge 1989, under which black bear populations currently exist. McLaughlin et al. 1994). In some cases, virtually Alternately, density effects may be undetectable due to complete reproductive failures followed extensive food methodological or circumstantial constraints. To failures (Jonkel and Cowan 1971, Pelton 1989). adequately assess the effects of density on population Physical condition and consequent reproductive ability dynamics, an assortment of population parameters apparently vary both temporally and geographically in should be measured under varying densities. However, response to food. Regional differences in reproduction densities of most black bear populations are too stable have been linked to differences in female weights and and methods of population estimation too imprecise to presumed or documented differences in forage even detect year to year changes in abundance, no less availability (Jonkel and Cowan 1971; Reynolds and relate such changes to subtle variations in birth and Beecham 1980; Bunnell and Tait 1981; Alt 1981, 1989; death rates, measured with equal imprecision. Kolenosky 1990; Stringham 1990b; Schwartz and A more fruitful approach might be to compare Franzmann 1991) (but see Beck 1991 for an alternate popUlation characteristics from a diverse array of study view). areas. Two principal drawbacks of this approach are: Some studies suggested that social interactions with (1) other conditions like food, weather, habitat, hunting adult males may preclude females from certain, pressure, etc. also differ among study areas, and may seemingly preferred habitats like rich blueberry patches affect population vital rates independent of bear density, (Jonkel and Cowan 1971), oak stands (Garshelis and and (2) methodologies vary among studies so some Pelton 1981), or even dumps (Rogers 1987a). Others perceived inter-study differences in population density noted possible sexual differences in habitat use, which and vital rates may be due to estimation error. may have been related to female avoidance of males Some population parameters, like cub litter size and (Hugie 1982, Pelchat and Ruff 1986, Clark 1991). If cub survival can be measured with little error, and so, it is conceivable that female nutrition may be partly methodologies are generally consistent across studies limited by the density of adult males, and reproduction (c.f., cub survival in Miller 1994). The procedures for may be consequently reduced where male densities are estimating inter-litter interval and age of first high. Similarly, because females in many areas are reproduction also are fairly standardized across studies, either territorial or at least temporally avoid parts of but the methods are biased toward short intervals and their home range occupied by other females (Reynolds young reproductive age, and the extent of bias increases and Beecham 1980; Garshelis and Pelton 1981; Hugie with increased rates of female mortality (Garshelis 1982; Pelchat and Ruff 1986; Rogers 1987a; Garshelis, 1993, Miller 1994). Additionally, some litter interval Minn. Dep. Nat. Resour., unpub!. data), it seems calculations include next-year replacement of whole­ equally tenable that increasingly high densities of litter loss, whereas most do not. 12 Int. Conf. Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

Estimates of population density and sex/age Z 7 0 composition are not only prone to significant error and 1= 0 0 bias, but also procedural inconsistencies. Although ~ a 6 0 0 0 most published studies employed some means of mark­ a: 0 recapture methodology, many investigators attempted to w0­ l a: 5 0 correct for inherent biases (due especially to differing I­ 0 0 f/) capture probabilities and movement of bears across a: 0 u:: 0 study area boundaries) using a variety of approaches LL 0 0 (Garshelis 1992). Moreover, some population estimates w included cubs, some did not, and in some reports it is ~ '1 0 31 unclear whether cubs were included. 10 20 30 40 With these caveats in mind, I compiled estimates of YEARLING WEIGHT (kg) density, sex/age composition, survival, and reproduction from black bear studies conducted Fig. 2. Significant inverse relationship between yearling throughout North America (Table 2). Many reports weights (at about 13-14 months of age) in 20 black bear provided only some of the sought-after parameter populations across North America and the average age that estimates, and in several cases I derived or revised females produced their first litter (data from Table 2). estimates from data presented by the author(s), which Populations with heavier yearlings likely had superior added yet another potential level of misinterpretation. nutritional conditions. so females matured earlier. Yearling weights were used as a surrogate measure of food abundance for each study. This age group was 1991; McLaughlin et al. 1994). chosen because in many black bear studies yearlings No relationship was observed between yearling were weighed in the den at 13-14 months of age, weights and litter size, inter-litterinterval, cub survival, thereby providing a standard for inter-study or yearling survival. Conversely, Stringham (1990b) comparisons. Weights of older bears may be a less found that the mean weight of adult bears from 7 comparable index of nutrition because: (1) ages of these hunted populations was a good predictor of both litter bears typically were estimated from dental annulations, size and inter-litter interval. Keay (1990) reported with potential error (Beck 1991, Coy and Garshelis reduced litter sizes in Yosemite National Park, 1992), (2) age groupings were inconsistent across California, after a reduction in availability of human studies, and (3) weights were measured at variable sources of food, but comparative studies within the times throughout the year, and therefore subject to states of Maine (McLaughlin et al. 1994) and Alaska significant seasonal effects. For the analyses herein, (Miller 1994) were unable to detect dissimilarities in weights of yearling males and females were combined litter size among study areas with greatly differing food because at this age they are often not appreciably supplies. Moreover, in Maine, McLaughlin et al. different (and sample sizes were thereby enhanced and (1994) found that inter-litter interval following whole analyses simplified). litter loss varied among their 3 study areas, but the A significant inverse relationship was observed interval between successful weanings did not. Miller between yearling weights and the average age of first (1994) observed that cub survivorship varied more reproduction (r = -0.78, n = 20, P = 0.0005) among years than among 3 Alaskan study areas. (Fig. 2). Noyce and Garshelis (1994) found that of all Similarly, Alt (1984) noted that year-to-year differences reproductive parameters, the age of primiparity of in weather conditions significantly affected mortality of females in northern Minnesota was most sensitive to cubs (due to flooding of natal dens), even for the variation in their weights, and thus proposed a heaviest black bears on the continent (A It 1980). Thus, generalized life-history model in which declining food although all reproductive parameters as well as cub and resources lead initially to older ages of first possibly yearling survival are undoubtedly related to reproduction. This model appears to be corroborated nutrition, these associations may be confounded by by the relationship observed here, using data from 20 other factors. Additionally, my analyses of these study sites across North America; yearling weights associations may have been hampered by using a reflected age of primiparity because both were surrogate variable like yearling weight to represent responsive to prevailing food conditions. Others also average nutrition; extremes in food abundance may reported a link between black bear nutrition and affect overall recruitment at least as much as average maturation (Beecham 1980a,b; Rogers 1987a,b; Beck levels. DENSITY EFFECTS ON BEARS • Taylor et at. 13

There were also no evident relationships between any likely less accurate (if based on mark-recapture of cubs) measure of recruitment (fecundity and cub survival) and than density estimates for older bears. Furthermore, if population density, adult male density, the percentage this inverse relationship was true, it would contradict of adult males in the population, or the ratio of adult previous observations suggesting higher cub mortality males to adult females (Table 2, Fig. 3). That is not to in years with relatively poor cub production (i.e., a say that none of these factors affected recruitment, but high percent of barren females) (Lindzey and Meslow rather if they did, the relationships were too complex to 1977a; LeCount 1982, 1990). My suspicion is that the be resolved by the data set compiled here. Based on weak relationship between cub survival and cub density previously presented theory, reduced adult male in the data compiled here was simply spurious. survival rates should correspond with increased In other cases I noted significant relationships that replacement by new males and intensified infanticidal further investigation revealed as spurious. For tendencies, but like the previous analyses, the cub example, from the data in Table 2 it appeared that bears survival rates compiled here did not relate to adult male produced larger litters in hunted populations survival (or the survival of any other sex-age group, (P < 0.05), possibly suggesting a compensatory although only adult males are shown in Table 2). response to higher mortality, although not to reduced Stringham (1985), using data from Rogers (1987a), density per se. Conversely, age of first reproduction noted correlations between cub survival and both male appeared to be delayed in hunted populations and female densities in northeastern Minnesota. (P < 0.05), which not only argues against However, it happened that food was scarce during years compensatory reproduction but also opposes expected with high bear densities, and Rogers (1987a) thought biases in the estimation of reproductive age due to that food was the predominant factor affecting cub mortality of females approaching maturity (Garshelis survival in his study. In the comparisons that I made 1993). Upon more careful examination of the data I using data from across North America, cub survival found that a preponderance of the unhunted populations was related (negatively) only to cub density (calculated occurred at more southerly latitudes. Previous authors by subtraction of densities including cubs minus those observed that black bears in eastern populations excluding cubs, Table 2) (r = -0.53, P = 0.02, (Minnesota and Arkansas eastward) tend to have larger n = 18). Likewise, in a study of northern fur seals litters than in the west (Alt 1989, Beck 1991); within (Callorhinus ursinus) (Fowler 1990), pup survival these 2 regions, my analyses indicated an increasing declined with increased pup density, but pup density litter size and increasing age of first reproduction from was considered an index of population density, not the south to north (ANOVA, P < 0.05). Additionally, agent affecting survivorship. It is puzzling how bear these data indicated significantly longer intervals cub survival could be related to cub density but not to between litters for western compared to eastern overall population density (Fig. 3) (or density of adult populations (Kruskal-Wallis test, P = 0.02). These males), especially since estimates of cub density were geographical effects seem to contribute more to reproductive variance than hunting, although the 100% reason(s) for these geographic clines remains elusive; 0 0 whereas nutritional differences may be at least one 0 80% underlying cause, the yearling weight data showed no @ 0 ....I 0 ~ 0 0 0 discernible geographic pattern (ANOVA, P > 0.1). 0 0 60% o 0 ;; § II: 0 0 :J (/) m 40% SUMMARY :J 0 0 u 0 Black bears have been studied across about a 20-fold 20% range of densities. It would seem that this variation

0% +-_---,--__,--_--,----_--,--__,--_---,-_---.J would be sufficient to discern density effects, if they o 20 40 60 80 100 120 140 exist. In a review of studies of other large mammals, (1981a,b, BEARS / 100 km' Fowler 1987) concluded that density­ dependent effects typically are exhibited only at population levels close to the environmental carrying Fig. 3. Insignificant relationship between bear density (excluding cubs) and cub survival in 21 black bear populations capacity. However, even at the highest black bear across North America (data from Table 2). Cub survival also densities, on Long Island, Washington, changes in was not related to density of adult males. reproduction, survival, and dispersal could not be 14 Int. Con! Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp. linked to the number of bears, but rather to diminished among studies (e.g., litter size, cub survival), but the food resources accompanying forest maturation. principal independent variables (e.g., density, sex/age Nevertheless, home ranges of bears on Long Island composition) were estimated with varying and uncertain (Lindzey and Meslow 1977b) were the smallest that I accuracy, or not at all (e.g., food abundance). Single­ am aware of, which may have been in response to the investigator comparisons of several areas seem exceptionally high density, and likely compounded the preferable in this regard, because methods can be impact of diminishing food availability. standardized. Clark's (1991) comparison of 2 areas in Effects of changing density on the vital rates of black the highlands of Arkansas demonstrated the efficacy of bears in a constant environment have yet to be this approach, and revealed convincing population observed. An increasing, but still moderately-dense regulatory effects (i.e., reduced cub survival and population on Stockton Island, Wisconsin, is currently reduced litter size). Nevertheless, the mechanisms for under investigation, but more years of study are the observed effects (e.g., male cannibalism of cubs and required to determine whether the seemingly increased consequent late reoccurrence of oestrus) were derived mortality there is density related. by conjecture. Moreover, even granting the Manipulative experiments provide a platform for hypothesized regulatory effects of adult males, this witnessing effects of selected population changes rather study showed only the potential consequences of an than waiting for such changes to occur naturally. uncommonly skewed adult sex ratio, not high bear Theoretically, then, the length of study can be shorter density. and confounding effects minimized. As I discussed Thus, none of the case studies, none of the data from relative to the Cold Lake study, however, manipulations various other studies, and none of the comparisons from of one population parameter (e.g., adult male density) the compilation of data among studies adequately could inadvertently alter others (e.g., probability of demonstrated density-dependent population effects in capture), which may introduce unanticipated and black bears. It does not necessarily follow that density­ possibly unmeasurable biases that affect perception of dependent relationships do not exist in black bear the population response. More structured experimental populations; however, presumed density effects in studies, though, have yielded more definitive evidence management programs certainly seem unwarranted, of density effects in other species (e.g., Bartmann et al. especially since they entail heightened demographic 1992). risks (Ginzburg et al. 1990). While there may be no Comparative studies circumvent artificially pressing need to incorporate density-dependent introduced biases from experimental manipulation, but components in black bear management, black bear do not escape error-driven inferences. In the continent­ researchers will and should continue to be intrigued by wide comparisons discussed here, some response the elusiveness of evidence in support of this variables were measured accurately and uniformly conceptually appealing regulatory process. DENSITY EFFECTS ON BEARS • Taylor et at. 15

DENSITY-DEPENDENT POPULATION REGULATION OF BROWN BEARS

BRUCE MCLELLAN, B.C. Ministry of Forests, RPO# 3, P.O. 9158, Revelstoke, B.C. VOE 3KO

If unaffected by humans or other predators, season. This behavior not only decreases the fitness of populations of large mammals in stable environments a competitor but may provide both food and an should be found close to carrying capacity (Fowler additional breeding opportunity for the male that did the 1981a,b). Factors suggesting that this concept may be killing. In response to infanticidal males, the particularly true for brown bears include: (1) large promiscuous mating behavior of females may have bodies that enable energy to be stored as a buffer to evol.ved to confuse paternity as has been suggested for short-term fluctuations in food supply, (2) hibernation several primates (Blaffer-Hrdy 1977). On empirical enables brown bears to avoid the consequences of grounds, adult males have been seen killing cubs, adult severe winters, (3) being omnivores, brown bears females with cubs often avoid habitats frequented by reduce the effect of single species food failures, and (4) adult males (Pearson 1975, Russell et al. 1979, Mattson brown bears are rarely preyed upon so are not affected et al. 1987), and a negative relationship between by fluctuations at a higher trophic level. Consequently, recruitment and the number of adult males in theory suggests that brown bears are expected to show Yellowstone National Park has been reported density effects when near carrying capacity (Fowler (McCullough 1981, Stringham 1983). 1981a). In reality, however, human influences may A literature review revealed 57 witnessed or rarely permit brown bear populations from attaining strongly suspected accounts of 1 brown bear killing these levels. another in a natural situation (Table 3). Of the 27 cases Factors that regulate populations can be categorized where the age and sex of the killer was known, all were as extrinsic or intrinsic. Extrinsic factors are external adults and 21 (78%) of these were male. Victims were to the organism such as resources, predators, and most often cubs-of-the-year (44 %), but adult females parasites, whereas intrinsic factors are characteristics of were also commonly killed (18%). Some, but not all the individual such as behavior, physiology, and genetic of these females were killed while protecting their cubs. potential (Sinclair 1989). These factors may operate alone or simultaneously to change vital rates or Table 3. Victims and known or suspected (in parenthesisl influence dispersal. killers in cases of intraspecific killing of brown bears in North As with black bears, an intrinsic factor, America. intraspecific aggression resulting in death or increased dispersal, is a potential mechanism of population Killers regulation. To be a regulating factor, the rate of Adult Adult Killer Victims females intraspecific killing and/or its influence on dispersal males unknown Total must increase with density. For terrestrial mammals in Adult females 0 6 (3) 2 12 general, however, the overwhelming cause of regulation Adult males 0 2 3 5 is intraspecific competition over food (Sinclair 1989) Subadults 0 3 (1) 5 resulting in reduced recruitment. In this chapter I review and summarize information related to the natural Cubs 5 5 (4) 11 25 regulation of brown bears in North America. Yearlings 4 2 7 Two-year-olds 0 1 (2) 2 5

DOCUMENTATION OF Total 6 21 (10) 20 57 INTRASPECIFIC KILLING .. References: Troyer and Hensel (1962), Pearson (1975), Reynolds On occasion, brown bears kill one another, and in (1976), Egbert and Stokes (1976), Glenn et al. (1976), Reynolds and particular, adult males kill cubs. This behavior may be Hechtel (1980,1984), Murie (1981), Nagy etal. (1983a,b), Dean et rare and insignificant, however, there is evidence, both al. (1986), Craighead et al. (1988), Miller (1990a,b), Reynolds and theoretical and empirical, suggesting that it may be Boudreau (1990), Smith and Van Daele (1991), Mattson et al. (1991), Olson (1993), Hessing and Aumiller (1994), Aune (Mont. Dep. Fish, important. On theoretical grounds, it could be an Wildlife, Parks, pers. cornmun.), Weilgus (Univ. British Columbia, evolutionary stable strategy (ESS) that adult males kill Vancouver, pers. cornmun.), Mattson (Univ. ofldaho, Moscow, pers. unrelated cubs, particularly prior to the breeding cornmun.), McLellan (British Columbia Forest Serv., unpub. data). 16 Int. Con! Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

Victims were of all age and sex classes, indicating that area may not result in an influx of young males and intraspecific killing is not limited to infanticide. thus increased infanticide may not occur. Although adult males were known or suspected of Studies reporting the possibility of density effects are killing cubs more often than adult females were, of the generally of 2 types. One type compares data from one 10 cases when the killer of cubs was known, it was a area before and after a population change was induced female 5 (50%) times. As suggested by LeCount by a change in management. The second type (1987a), females could also benefit from killing the compares 2 areas of similar habitat but with different cubs of unrelated females. Could the lack of dispersal management goals and thus different population density by females be an ESS that reduces infanticide by or sex and age structure. unrelated females? Possibly, but because 3 of the 5 cubs killed by adult females were killed in 2 separate Yellowstone: 1959 to 1970 instances by 1 individual (Hessing and Aumiller 1994), Between 1959 and 1970, the Craighead study team cub killing by females may be considered abnormal collected an exceptional data set on the brown bears of behavior at this time. Yellowstone National Park (Craighead et al. 1974). The number of observed cases of intraspecific killing Due to the duration of study, large annual sample sizes, may have several significant biases. Observations were plus the controversy over bear management in most often made either in open habitats where cubs Yellowstone, these data have been re-analyzed in great could not climb trees to escape, or where bears detail (Schaffer 1978, 1983; Stringham 1980, 1983, concentrated, promoting frequent interactions. On the 1985, 1986; McCullough 1981, 1986). Although the other hand, few observations were made before Yellowstone data do not fit either a "before and after" breeding season when a disproportionate amount of cub or "2 study area comparison" design, these authors mortality occurs (Reynolds and Hetchtel 1984). The noted a negative relationship between reproductive direct effect of intraspecific killing on population parameters and adult population size or number of adult processes remains equivocal; however, when combined males and concluded that a density-dependent effect was with the indirect effect that this aggression may have on being observed. dispersal rates and segregated habitat use, intraspecific I will not conduct another detailed analysis of the killing may be very significant. Craighead data. However, I feel compelled to review certain aspects of previous reports, especially given McCullough's (1981:177) assertion that "the negative CASE STUDIES WITH correlation between recruitment of young and adult REPORTED DENSITY EFFECTS population size demonstrates beyond doubt that the Population manipulations with a species such as the Craighead et al. (1974) data indicate a very strong brown bear are difficult to justify for purely scientific compensatory factor. " reasons, and so far, none have been conducted. The Craighead team collected population data mostly Researchers have, however, used various management at garbage dumps where bears concentrated. Between programs as experimental treatments. When 1964 and 1970, the number of adult females censused investigating population regulation, large scale but less varied little (range 40 to 48) whereas the number of controlled management programs may be preferred to adult males varied greatly (range 27 to 53). Bears were smaller scale, highly controlled experimental not categorized by sex before 1964, but the number of manipulations. For example, to test the hypothesis that adults of each sex was estimated by McCullough (1981) populations are regulated by adult males killing cubs, a and Stringham (1983) from the relationship between the researcher might remove adult males and monitor number of adult males and total number of adults changes in the rate of infanticide. In many mammals, between 1964 and 1970. Combining the 5 years when males that kill infants are usually not their fathers, but adult male abundance was estimated with the 7 years are often immigrants who replace resident males that it was recorded, indicated that the number of (Hausfater and Blaffer-Hrdy 1985). If this behavior different males visiting dumps varied 3-fold (18 to 55 was also typical of brown bears, a small-scale individuals) over the 12-year study (Fig. 4). experimental removal could encourage settlement by Annual changes in total adults averaged 14% (range young males who were not fathers of the resident cubs o to 25 %) whereas the annual change in adult males and thus increase infanticide (Stringham 1980). A averaged 27% (range 7 to 61 %). Why did the number management-scale experiment using hunting to reduce of males fluctuate so much? Natural survival rates of the number of adult and subadult males over a large adult brown bears is high and consistent (Craighead et -- DENSITY EFFECTS ON BEARS • Taylor et al. 17

60

50 en a: c( w III 40 ~ ::::» Q c( 30 II.o a: w 20 III :::t ::::»z 10

• FEMALES o o MALES 59. 80. 81. 82. 83. 84. 85. 88. 87. 88. 89. 70. YEAR

Fig. 4. The number of adult bears of each sex recorded by Craighead et a!. 1974 in Yellowstone between 1959-1970. al. 1974, Knight and Eberhardt 1985, McLellan 1989a) McCullough's (1981) analysis of the dynamics of the and changes in adult male numbers visiting the dumps Yellowstone population was based on regressions of did not correlate with recorded mortality in the recruitment, measured as cubs per adult, on the number ecosystem (r = 0.09, P = 0.79). Similarly, the of adults. He suggested strong density dependence reproductive rate of brown bears is low so even highly because he found the number of cubs as well as skewed sex ratios at birth could not explain the yearly yearlings and 2-year-olds per adult to significantly variations in sex ratios observed among adults. decrease (r = 0.84; P < 0.0005 for cubs) as the I suggest that the censuses at the Yellowstone dumps number of adults increased. were not an accurate reflection of annual changes in the McCullough's (1981) analysis is flawed in 2 ways. population size of each age/sex class in the ecosystem. First, as noted by Stringham (1983) but defended by One explanation is that during years of low natural food McCullough (1986), was that the regression abundance the number of wide-ranging adult males (cubs/adults versus adults) has the same variable on visiting dumps increased dramatically and both axes and this could produce a negative relationship disproportionately to females while in good food years from auto-correlation alone (Stringham 1983). For males spent less time there. Recent research in example, when the recorded adult population was Yellowstone indicates great annual fluctuation in natural higher than average but the cub population was near food abundance and an inverse relationship between average, cubs per adult would be low; conversely, food production and bear management action (Knight et when adults were low and cubs near average, cubs per al. 1988a, Mattson et al. 1991, Mattson and Reid adult would be high. I investigated this problem further 1991). Increased dump use by radio-collared black by generating 20 normally distributed but random data bear males during poor food years was clearly apparent sets with the same mean and variance as the in Minnesota (D.L. Garshelis, Minn. Dep. Nat. Yellowstone data. Only 2 of 20 random data sets failed Resour., pers. commun.). to show a significant relationship and 3 were not only 18 Int. Con! Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp. significant but had r values higher than the 0.71 and Knight 1986). Additionally, this climate index only McCullough (1981) concluded were density dependent. explained 37 % of the variation in litter size during the A recent review of the problem of autocorrelation in years of open dumps. studies attempting to show density effects was given by Some biologists interpreted the analyses of Messier (1991). McCullough (1981) and Stringham (1983) as The second difficulty with the analysis of demonstrating that adult males could have suppressed McCullough (1981) is that the change in adults was recruitment by killing cubs and thus justify the hunting largely caused by just a change in the number of adult of males in small populations (e.g., Dood et al. 1986). males. Thus the negative relationship between cubs per If this hypothesis is correct, one would expect that the adult and adults may only show that males do not have survival rate of cubs (measured as the difference cubs. It may be more meaningful to regress cubs per between the number of cubs censused one year and adult female against the total number of bears recorded yearlings the next), to be inversely related to the in the censuses. This relationship was not significant number of adult males. However, in the Yellowstone (P = 0.72). data, neither cub or yearling survival were related to Stringham (1983:140) conducted a different analysis the number of males or total adults. of the same data set and concluded that "years when The intent of this discussion is to indicate that density adult males were most abundant were also those in effects are not unambiguously demonstrated in the which: 1) the litters conceived were smallest when Craighead's Yellowstone data. Additional information censused at median age 0.5 year postpartum, and 2) the (e.g., average number of adult males in each census cohorts born were comprised of fewest litters at that rather than the total over the year plus the occurrence age"; he offered 2 possible explanations. First, adult of known individuals among years) may strengthen or males control reproductive rates and recruitment by weaken competing hypotheses. It is unlikely, however, infanticide, exile, resource competition, or socially that these intriguing data will ever be fully understood. induced stress. Second, adult male abundance may In my opinion, it will be difficult to identify density have been merely an index of other factors such as food effects of a "K" selected species by studying one supply. The significant inverse relationship between population over time while relying on natural changes cub production and the number of males seen in dumps in population density. Although density effects may be the year prior to birth may support the second strong, most between-year variation in vital rates of a hypothesis more than the first. If adult males were population as it tracks a gradually fluctuating carrying most abundant in dumps during poor food years then capacity would be caused by density-independent females in these years would be in relatively poor events. Manipulations that change population condition from a lack of natural food and added characteristics over the range of densities where competition in the dumps. Consequently, the following density-dependent regulation is experienced is likely spring, litters would be small and uncommon. required for robust evidence of density effects (Sinclair Stringham (1983:145) found "litter size was more 1989, Appendix I, this paper). Miller (1990b) cautions strongly related to prenatal than to postnatal abundance that analyzing raw ecological data collected by other of adult males, whereas litter abundance was [inversely] people for other reasons must be done critically and related only to postnatal abundance of adult males." carefully. Even with well designed experiments, The reduced number of litters censused during years confounding factors are common. I conclude that the when adult males were most abundant in dumps could Craighead's Yellowstone data neither demonstrate nor be caused by infanticide. Alternatively, just as females refute the various hypotheses that have been advanced with cubs appear to avoid habitats used frequently by suggesting density-dependent population regulation. males (Pearson 1975, Russell et al. 1979, Mattsonet al. 1987), they may have avoided dumps when males were Susitna River, Alaska abundant there. Miller (1990b) monitored cub survival in a portion of Unfortunately, my interpretations of the Yellowstone his study area in the Susitna drainage, Alaska, that not situation is hampered by a lack of data on food only had a long history of hunting, but a recent, abundance. The climate index derived by Picton (1978) dramatic increase in hunting. Between 1974 and 1979, to predict litter sizes may not be an appropriate index an average of 65 bears were harvested per year. The of summer food abundance as it focused on the harvest was then increased to an average of 116 per perigestational period, between blastocyst implantation year during 1983 to 1988 (Miller 1990c). Although in the late autumn and den emergence in spring (Picton Miller (1990a) estimated his entire study area to have DENSITY EFFECTS ON BEARS • Taylor et al. 19 habitat of similar quality, he found portions with and mild weather conditions (Sellers and Aumiller historically high bear harvest to have about half the 1994). The conclusion of a lO-year ground-based density as a lightly hunted portion even prior to the capture operation with related disturbance and increased harvest. After the harvest increased, the mortalities (10 bears) in 1973 , and the end of density of bears in the heavily harvested area was unregulated public visitation may have resulted in an reduced to about one-third that of the more lightly additional increase of bears at the concentration area. hunted portion and both adult and subadult males were As was the case with the Yellowstone dumps, relatively rare (Miller 1990c). Miller (1990b) reported MRSGS is a seasonal concentration site and annual 4 of 6 deaths of radio-monitored cubs were due to censuses done at the falls may not mirror changes over predation by brown bears of unknown sex during his the entire population, particularly since some important study but he did not detect an increase in cub survival factors influencing bears were limited to the census rate with a decreasing population and a decreasing area. It is possible that the increase in bear numbers at proportion of males. Similarly, litter sizes averaged the concentration site was due primarily to a shift in 2.05 (n = 19) before the major increase in harvest rate areas used by some bears after human induced stress and 2.09 (n = 44) after. Weaning interval also did not around the falls was reduced. However, because sport respond to the change in bear numbers, averaging 3.2 hunting was curtailed or eliminated over a large area years (n = 11) before and 3.3 years (n = 19) after the near MRSGS and censuses elsewhere on the Alaskan increased harvest. More recent information suggests Peninsula also found populations to be increasing inter-weaning intervals may be increasing at reduced (Miller and Sellers 1990), it is probable that a true population levels (S. Miller, Alas. Dep. Fish and population increase was realized. With only 2 years of Game, pers. commun.). Clearly no compensation by data after the number of bears apparently peaked in increasing reproduction or decreasing mortality has yet 1989, it was not yet possible to say that the bears at the been noted. Miller (1990b:465-466) suggested that falls had reached carrying capacity. Additionally, the "bear density certainly influences productivity and number of bears using the concentration area may probably influences survivorship in high density bear plateau before the carrying capacity of the broader area populations existing at or near habitat carrying is reached. More information is required to both capacity" but he did not detect a density-dependent establish that this population is currently regulated by response in his study area where densities had density related factors, and clarify the biological previously been reduced (Miller 1990c). mechanism population regulation if it is occurring.

McNeil River, Alaska Glacier National Park, Montana Sellers and Aumiller (1994) reported 23 years of In Glacier National Park (GNP), Montana, brown population data from McNeil River State Game bears have not been hunted for many years, predator Sanctuary (MRSGS). The total number of bears control declined outside the park in the mid-1950s recorded at MRSGS was stable between 1969 and 1984 (Keating 1986), and control killing was light compared and averaged 66, but then increased to 126 bears by to the estimated population size (Martinka 1974). 1989 and then decreased slightly in 1990 and 1991. Consequently, the brown bear population is likely near The measurements of reproduction and recruitment used carrying capacity. Because litter sizes reported in 2 by Sellers and Aumiller (1994) did not decrease as the studies within GNP were among the smallest reported total numbers of bears or the number of adult males in North America, while 2 hunted populations increased. During the period of rapid population immediately outside the park were among the largest growth between 1983 and 1989, all segments of the (Aune et al. 1986, Mclellan 1989b); McLellan (1989b) population increased in number except for subadults. suggested that a density-dependent reduction in park This observation led Sellers and Aumiller (1994) to bear reproductive rates could be occurring. Mclellan speculate that subadult survival and/or dispersal may be (1989b) speculated that females in the park may be the density-dependent mechanism that limits growth of forced to use poorer habitats due to a relative the population at MRSGS. abundance of males that use preferred sites as has been Reasons for the increase in bear numbers in the suggested for bears in other unhunted parks (Pearson vicinity of MRSGS were unclear, but likely involved 1975, Russell et al. 1979, Mattson et al. 1987). Miller the expansion of hunting reserves to the south of the (1990b) correctly criticized this interpretation because sanctuary in 1978 and 1985, more restrictive hunting of possible habitat differences between the 2 areas in regulations to the north in 1975, and strong salmon runs the park and 2 areas outside the park. 20 Int. Con! Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

To further investigate the alternate hypotheses To investigate compensatory responses of brown causing different litter sizes (habitat differences versus bears to hunting, Stringham (1980) compared hunting) comparisons were made between park and non­ information on 6 populations. Like other comparisons, park areas in the same valley and at same elevations to his analyses were confounded by habitat quality. Bears ensure habitats were similar. The North Fork of the in good habitats were at higher densities and had higher Flathead River flows north to south through a wide, reproductive rates than bears in poorer habitats and flat-bottomed valley. At 49° north latitude, the valley consequently, he could only conclude that available data crosses the border and management changes. North of were inadequate to show density-dependent effects. the 49° parallel there was timber harvest and hunting, Nagy and Haroldson (1990) compared home-range and to the south there was the unhunted park. North of size, body size, and population characteristics among 4 the border the average litter size of radio-collared areas where data were collected in similar ways. One females was 2.2 cubs (n = 26) (McLellan 1989b). population, Northern Yukon, was virtually unhunted Sighting data from GNP files only from the north end and thought to be at carrying capacity. This population of the North Fork drainage revealed a mean litter size was at a much higher density, had older males and of 1. 7 (n = 48), the same as reported by Kendall smaller bears than the other populations. Although data (1985) and Martinka (1974) in other parts of the park were insufficient to adequately compare reproductive and significantly different from immediately north of parameters, Nagy and Haroldson (1990) speculated that the border (P = 0.001). Nevertheless, this does not reproduction was density dependent and that at high rule out effects of habitat, as successional differences densities, most females would only be able to produce due to timber harvest north of the border or some other cubs during years of abundant food production. Unlike factor may be significant. In addition, litter sizes in Blanchard (1987) and Stringham (1990a) who both GNP were derived from casual observations that may suggested that body weight was a reliable index of miss cubs more frequently than in the telemetry study habitat quality, Nagy and Haroldson (1990) stated that, outside the park. However, unhunted populations such in the absence of information on population density, as in GNP where bears should be close to carrying body weight could not be used to reflect habitat quality. capacity, should provide the best opportunity to observe I tabulated population data from 17 areas across density-dependent effects. A detailed ecological study brown bear range in North America (Table 4) with in GNP to compare to the research in adjacent areas reasonable sample sizes. Body weights were estimated would thus be very informative. by summing the average adult (> 5 yrs) \,\pril to July weight and the average August to November weight and dividing by 2. In populations where bears were COMPARISONS AMONG POPULATIONS weighed only in spring, the mean spring weights were Everything else being constant, reproduction and/or multiplied by 1.28 for females and 1.24 for males (the survival rates should decrease when we compare low­ average seasonal increase exhibited in populations to high- density populations. As discussed in the where bears were weighed in both seasons). Due to previous section, everything else never remains previously mentioned problems regarding the constant. Even comparisons of populations of brown consistency of data collection, detailed statistical bears living in adjacent portions of the same valley may analyses cannot be justified for most variables. My be confounded by differences of habitat quality. This analyses were conducted for reconnaissance purposes, problem is even more exaggerated when comparing data and the reader is advised to view the results with from across the continent, where recorded densities caution. vary by almost 2 orders of magnitude; moreover, These data support some of the finding of previous because most populations are exploited in some way, authors. In particular, coastal populations have larger most are probably not near carrying capacity. bears living at much higher densities (Table 4). One Comparisons are further hampered by differences in the explanation that is consistent with these observations is methods used to estimate population parameters. that coastal areas produce more bear food than interior Density, inter-litter (sometimes interweaning) interval, areas. However, habitat quality has not been quantified and adult sex ratio were often defined and estimated in any study area; so these data support but do not differently. Additionally, there may be genetic demonstrate intraspecific competition for food as a differences between populations that cause population general regulating factor. Some of the very high characteristics to differ either independent or in densities recorded from coastal areas may also reflect response to density. limited dispersal opportunities from islands or the DENSITY EFFECTS ON BEARS • Taylor et al. 21

Table 4. Estimated characteristics of grizzly bear populations in north America with sample sizes in parenthesis. Some variables have been collected in different ways among studies so these data must be used cautiously.

\ ( Density] Cub Reproductive Age at Adult Adult Cub % Adult Study area bearsll 00 km2 litter size interval l first litter male weight female weight mort. rale male l Hunted? INTERIOR POPULATIONS East Front Montanaa,b 0.7 2.2 (41) 2.6 (11) 6.0 (4) 125 54 yes Flatheadc 8.0 2.2 (26) 3.1 (17) 6. I (7) 176 (22) 114 (16) 0.18 37 yes Eastern Brooksd 0.4 1.8 (13) 179 (26) 108 (31) 49 yes Alaska Rangee,f 1.53 2.2 (36) 4.22 (38) 7.62 (8) 2244 (24) 1354 (32) 0.29 33 yes

Nelchinag 1.0 2.1 (64) 3.82 (44) 5.62 (24) 2694 (12) 1444 (21) 0.30 275 yes Tuktoyaktukh 0.4 2.3 (18) 3.32 (8) 6.42 (10) 195 (16) 124 (36) 33 yes MacKenzie Mountainsi 1.2 1.8 (6) 3.8 (5) 148 (20) 110 (28) yes Glac ier Parkj 4.7 1.7 (35) no

Yellowstone 1959_70k,I,m,n 2.2 (173) 3.2 (68) 5.7 (16) 245 (33) 152 (72) 0.26 46 no

Yellowstone 1975-89n,0 1.9 (232) 2.6 (20) 5.7 (23) 193 (65) 134 (63) 0.15 55 no Western BrooksP 2.4 2.0 (57) 4.1 2 (16) 7.92 (14) 182 (26) 117 (35) 0.44 42 no K1uane Parkq 3.7 1.7 (11) 7.7 (7) 145 (26) 98 (16) no Northern Yukonr 2.8 2.0 (6) 4.02 (4) 7.02 (3) 173 (59) 116 (35) 51 no COASTAL POPULATIONS Kodiak Islands,t 28.0 2.5 (29) 4.62 (41) 6.72 (12) 312 (10) 202 (16) 0.37 38 yes Alaska Peninsulau,m 18.4 2.3 (200) 3.0 (81) 4.4 (9) 357 (21) 226 (63) 0.40 28 yes Admiralty Islandv 40.0 1.8 (32) 3.9 (7) 8.1 (7) 2604 (10) 169 (18) 0.20 yes McNeil Sanctuaryw,m 2.2 (137) 3.8 (37) 6.5 (11 ) 257 160 0.31 55 no a Aune and Brannon (1987) Craighead et al. (1976) b Aune and Kasworm (1989) m Stringham (1990a) c McLellan (1989a,b,c; British Columbia For. Servo unpubl. data) n Blanchard (1987) d Reynolds (1976) o Knight et al. (1988, 1989, 1990) e Reynolds (1989) p Reynolds and Hechtel (1984) f Reynolds and Boudreau (1990) q Pearson (1975) g Miller (1990a,b,c) r Nagy et al. (l983a) h Nagy et al. (1983b) s Smith and Van Daele (1991) Miller et al. (1982) t LeFranc et al. (1987) j Martinka (1974) u Miller and Sellers (1990) k Craighead et al. (1974) v Schoen and Beier (1990) w Sellers and Aumiller (1994) 1 Due to a variety of methods used in their derivation, comparisons must be done cautiously. 2 Includes incomplete or potential intervals and births. 3 Adjusted from bears> yearlings to all bears by 1.3 (Miller 1988). 4 Spring only weights and adjusted by 1.28 for females and 1.24 for males. 5 Adult sex ratio changed from 53 % to 27 % male during study period due heavy harvest.

Alaskan peninsula, suggesting the possibility of a "fence coastal populations were excluded from the analyses. effect" (Krebs et al. 1973). At this time, however, Similarly, cub survivorship was not related to bear dispersal behavior and its relationship to bear density is density (r = 0.1, P = 0.81), the proportion of adults poorly documented. that were male (r = 0.37, P = 0.32), nor whether the Using information from 14 studies with appropriate population was hunted or not (P = 0.92). It appears data, none of the reproductive parameters were that the great variation in habitat quality across brown negatively correlated with bear density, even when the bear range precludes simple comparisons. In fact, r

22 Int. Conf. Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

relationships among habitat quality, body size, and Using a mark-recapture technique, Schoen and Beier reproductive parameters are even less clear than earlier (1990) estimated a density of 40 bears/1 00 km2 on their investigators reported. For example, based on data study site on northern Admiralty Island. Extrapolating from 8 studies, Blanchard (1987) found a high their density estimate to the entire island results in a correlation between female body weight and litter size population of 1761 bears. Using this figure, the (r = 0.92). Using information from 7 of these areas average harvest of 42 bears between 1984 and 1988 (Jasper Park data were not used because of small (Young 1990) represents a 2 % harvest rate for the sample sizes) and 9 other studies, the relationship island as a whole. The overall harvest level is remained significant but the correlation between female considered well below the sustainable yield. body weight and litter size was considerably reduced Typical of the coast, bears were large on Admiralty (r = 0.59) (Fig. 5). and the density was the highest so far recorded Bears in several study areas clearly deviate (large anywhere except at seasonal concentration sites. residuals) from the body weight to litter size However, unlike other coastal areas, females on relationship and these anomalous areas may provide Admiralty mature late, have long inter-litter intervals, information on the potential for cub production to and small litters (Table 4). regulate populations of brown bears (Fig. 5). The area Schoen and Beier (1990) suggested that the high with the largest negative residual was Admiralty Island density of male and female bears on Admiralty Island (residual = - 0.38) while large positive residuals were caused adult predation on cubs to contribute to the small from Tuktoyaktuk (residual = 0.29), the Flathead litter sizes. However, annual cub mortality (20%) was (residual = 0.22), and Kodiak Island (residual = 0.20). relatively low. The high density of bears on Admiralty Admiralty Island is hunted and the harvest has appears to have somehow affected the reproductive recently increased, particularly in accessible areas. rates by delaying maturation, reducing litter size, and increasing the interval between litters. General observations suggest that the bears were not having an 2.8 impact on the food supply and thus food itself was , KI unlikely the proximate regulating factor (J. Schoen, Alas. Dep. Fish and Game, pers. commun.). 2.4 However, social stress was clearly evident (Schoen pers. commun.), and may have resulted in forcing III ro. .U ~ subordinate bears into suboptimal habitats, increased U) a: 2.2 energy expenditure, plus increased vigilance and thus III inefficient foraging. On Admiralty, the distribution of ... N food and thus suitable feeding sites may be the ultimate 5 NY 2.0 WB' regulating factor while social behavior may be the Y2. proximate mechanism. !2 MM. Other anomalies of the body weight to litter size .AI 1.8 R. regression include bears from Tuktoyaktuk, the KP. Flathead drainage, and Kodiak Island. The first 2 are interior populations with small bears but large litters 1.8 and moderate inter-litter intervals. Kodiak Island is on SO 100 150 200 250 the coast and had large bears and very large litters. All MEAN ADULT FEMALE WEIGHT (KG) 3 populations have a history of high human-caused mortality, particularly by hunting. Although Nagy and Haroldson (1990) suggested that Fig. 5, The mean litter size of cubs and the mean weight of adult females is given for various populations of brown bears. the Northern Yukon population was at carrying The populations are identified as follows: East Front lEF), capacity, the bears' body weight and litter size did not Flathead (FH), East Brooks (EB), Alaska Range (ARl, Nelchina deviate from the regression line (residual = 0.02) as (NEl. Tuktoyaktuk (TU), MacKenzie Mountains (Mel, did the data from Admiralty Island. This may be due Yellowstone 1959-1970 (Y1l, Yellowstone 1975-1989 (Y2l, to a small sample of litters (n 6) and/or density Western Brooks (WBl, Kluane Park (KPl, Northern Yukon (NYl, = Kodiak Island (KIl, Alaska Peninsula (APl, McNiel River (MRl, dependence was being expressed in an extended inter­ and Admiralty Island (All. The data for this figure are litter interval; with 54 % of the females being alone in contained in Table 3. the spring (Nagy et al. 1983a), the real inter-litter DENSITY EFFECTS ON BEARS • TayLor et aL. 23

Table 5. Percentage of breeding age females, including the year prior to their first litter. that were alone, with cubs. yearlings. 2. 3, or 4-year aids in the spring.

Study area % Alone % Cubs % One % Two % Three % Four N

Tuktoyaktuk 15 33 26 22 5 0 59

North Yukon 54 16 15 12 2 2 46

Flathead 16 38 30 16 0 0 63

West Brooks 30 28 17 16 7 123

Alaska Range 15 32 27 20 6 0 133

Admiralty 35 25 18 13 7 127

Terror Lake 31 22 26 17 5 0 143

McNeila 49 23 28 b 357 a Data from 1976 to 1991 b Includes all juveniles > 1 year old with their mother. interval must have been longer than the 4 years populations appear complex and I can only speculate on calculated from the sample size of 4 intervals. their relationships. Because overall densities vary Schoen and Beier (1990) mentioned an important bias greatly among areas and appear to be related to food in estimating inter-litter (or inter-weaning) interval productivity, food is likely the ultimate regulating observed on Admiralty Island. Several of their radio­ factor. However, in most systems, bear consumption marked bears went several years without having cubs, does not appear to reduce food biomass to a level where past the duration of radiotracking. Because an interval foraging efficiency is impaired. In systems where food for those females was never measured, they were not is clumped, foraging efficiency is likely impaired at included in the interval calculation. This appears to be high densities by social behavior causing displacement a common bias. Table 5 reports the reproductive status from feeding sites, increased vigilance, and increased of females prior to weaning in the spring for energy expenditure due to social stress. Intraspecific populations for which these data were available. killing also appears important, particularly where there Females in heavily hunted populations such as is little security and escape cover. In systems where Tuktoyaktuk, Flathead, and the Alaska range were food is more evenly distributed, food depletion at high found without cubs less frequently than the unhunted or densities may be more significant. lightly hunted populations in the Northern Yukon, Although much has been published on reproductive McNeil River, or Admiralty Island. rates and more has recently been reported on mortality rates, almost nothing has been published on dispersal. Because food production and distribution appear to SUMMARY interact with social behavior as regulating factors, a Only a few areas had lightly hunted brown bear social "fence effect," as described for some small populations that were likely near carrying capacity. mammals (Krebs et al. 1973, Hestbeck 1982) may These few (i.e., Admiralty Island, McNeil River State operate. In such a scenario, brown bear populations Game Sanctuary, and the northern Yukon) gave some would be regulated at one level by socially-induced indication that density-dependent effects may have been dispersal; reproduction and survivorship may be high in operating. Bears on Admiralty Island and the northern the core area and reduced or zero at surrounding Yukon appeared to have reduced reproduction, and in "population sinks." Alternatively, a higher equilibrium particular, long inter-litter intervals. Bears at McNeil density may occur where dispersal opportunities are River also had long inter-litter intervals, but it was limited because of high densities over a large area or by subadult dispersal that was suggested to be the geographical boundaries. In such an area recruitment mechanism of regulation. and/or survival will be reduced in the core area. Factors affecting the dynamics of brown bears These hypothesis of the biological mechanism for - 24 Int. Con! Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

These hypothesis of the biological mechanism for Understanding the mechanism of density-dependen density-dependent population regulation are not population regulation in a brown bears population wil mutually exclusive or conclusively demonstrated for any require intensive monitoring and a long-term researct population. The prudent manager should not assume effort. These efforts will have to be repeated ir that a reduction in density will cause an increase in the different geographic areas before the generality of an) recruitment or survival of a given brown bear conclusions is apparent. population, especially one that is already harvested. DENSITY EFFECTS ON BEARS • Taylor et al. 25

DENSITY-DEPENDENT POPULATION REGULATION OF POLAR BEARS

ANDREW E. DEROCHER, Canadian Wildlife Service, 5320 122nd Street, Edmonton, Alberta T6H 3S5 MITCHELL TAYLOR, Department of Renewable Resources, 600, 5102 50 Avenue, Yellowknife, Northwest Territories X1 A 3S8

Systematic studies of density effects on polar bears Pregnant females undergo an 8-month fast, during have not been undertaken. Part of the problem is that which they give birth to and nurse cubs (Watts and most studied populations have been reduced by harvest. Hansen 1987, Ramsay and Stirling 1988). If fat stores Harvested populations may have been reduced to are insufficient, pregnant females may not produce densities at which density effects were insignificant. cubs, and females with offspring may not be able to Another problem is that the polar bears inhabit remote nurse them. Derocher et al. (1993) document changes arctic areas where it is difficult to study them. in polar bear milk composition associated with fasting. Expanding a polar bear study to include prey species Pregnant females with lower body weight are less and social interactions is logistically difficult and very successful at rearing cubs possibly because of increased expensive. Although no study has claimed to have cub mortality associated with lactation failure. Thus the demonstrated density effects in polar bears, some biological mechanisms of nutritional density effects authors have speculated on possible mechanisms. would probably not significantly alter adult survival, Recognizing that the available data are not sufficient to rather it would involve cub production and cub confirm or refute the existing hypothesis, we review the survival. supporting evidence. Polar bear populations might also be regulated by Population regulation could be mediated by either social factors, including but not limited to actual nutritional or social factors. Most likely, the effects of conspecific killing. The average interval between litters density would be both nutritional and social, as is 3.7 years (Taylor et al. 1987a) which means that suggested (this monograph) for both black and brown mating opportunities are scarce for adult males (Ramsay bears. Nutritional regulating factors would result in and Stirling 1986). If males could locate females with polar bear populations that were maintained within cubs during the mating season, particularly newborn bounds determined by variation in the density and cubs, infanticide followed by copulation might induce availability of the major prey species, the ringed seal ovulation. Alternatively, or additionally, large males (Phoca hispida) (Stirling and Archibald 1977, Smith might prey on other sex and age groups if encounters and Stirling 1978, Hammill and Smith 1991). Like were common. polar bears, ringed seals are long-lived marine The nutritional and social hypotheses are not mammals whose populations do not fluctuate greatly mutually exclusive. Subdominant sex and age classes over short time intervals. However, annual fluctuations could be excluded from important feeding areas by in production of ringed seal pups, an important food dominant animals during the hyperphagic period (April, source, do occur (Smith and Stirling 1978). The May, June). No polar bear study to date has occurred hunting success of polar bears for ringed seals is over a time span sufficient to quantify changes in probably also influenced by ice conditions and weather popUlation numbers and changes in rates of birth and pattems, which can vary considerably between years death. However, studies from western Hudson Bay, (Dunbar 1981, Smith and Rigby 1981, Stirling et al. the eastern Beaufort Sea, and Svalbard do provide some 1988, unpubl. data) insight into potential biological mechanisms by which During hyperphagic periods (particularly seal density could affect polar bear population dynamics. pupping and seal molting), polar bears deposit fat reserves used during fasts when prey are less available. Polar bears, like black and brown bears, are CASE STUDIES WITH physiologically and behaviorially well adapted to such POSSIBLE DENSITY EFFECTS fasts (Lunn and Stirling 1985, Watts and Hansen 1987, Ramsay and Stirling 1988, Derocher and Stirling 1990, Western Hudson Bay (Churchill, Manitoba) Derocher et al. 1990), but if insufficient fat reserves are Research on the popUlation of polar bears that available, reproductive success and cub survival are summer near Churchill, Manitoba has occurred reduced (Derocher 1991, Derocher and Stirling 1992). continuously each year since 1966 (Jonkel 1967, ...... 26 Int. Con! Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

Stirlinget al. 1977b, Derocher and Stirling 1992). The effect, even though the densities of polar bears hav recruitment rates noted for the Churchill population in remained constant. If a population is at (or close to the early 1980s were the highest found for this species carrying capacity, and the carrying capacity is reduced in any population. However, during the period (1980­ the population will respond by declining to the ne, 1990), the mean weight of polar bears in western (lower) carrying capacity. The Hudson Bay-James Ba: Hudson Bay declined (Derocher 1991, Derocher and populations are the most southern of all polar bea Stirling 1992). Presumably in response to the decline populations, and may be more sensitive to variation il in condition, cub survival in the first 6 months after den nutrition because of fasting during the extended ice-frel emergence declined from 75% in the early 1980s to season. 51 % by the late 1980s; and death of whole litters The fourth possibility (Derocher 1991, Derocher an< increased 3-fold (Derocher 1991, Derocher and Stirling Stirling 1992) is that the current densities of polar bear: 1992). These changes coincided with closure of the may have nothing to do with the observed declines il Churchill dump to polar bears; however, only an maternal weight, cub surviVal, and recruitment. Thl insignificant number of the Churchill population used observed reduction in female condition, cub production the dump, and that use was determined to be and cub survival may be simply a density-independen nutritionally insignificant (Lunn and Stirling 1985). response to a run of difficult years that happened t< These dump bears were excluded from the analysis follow a period of better conditions. The available dat< showing declining condition and cub survival (Derocher are inadequate to resolve these possibilities. and Stirling 1992). Derocher and Stirling (1992) No evidence for increased intraspecific aggressior offered 4 explanations for the declining weight of has been presented. However, this population Wal Churchill bears. The first 2 hypotheses concern the per studied onshore during the post-mating summer retreat capita availability of seals to bears: (1) a population Spatial segregation by family group and sex has beer increase that resulted after a long period of unregulated observed in the Hudson Bay population in fall (LatoUJ harvest and/or (2) a decline in the abundance of ringed 1981, Derocher and Stirling 1990) and othel seals. However, the density of polar bears was not populations in spring (Ramsay and Stirling 1986, observed to change during the 1980-91 period Stirlinget al. 1993, unpubl. data). Although emigration (Derocher 1991). There is no information on ringed from the western Hudson Bay population appears to be seal abundance in Hudson Bay. very limited (Derocher and Stirling 1990), large males The third possibility (Derocher 1991, Derocher and might have excluded females with cubs and other Stirling 1992) was that ice conditions adversely affected smaller bears from prime feeding habitats on the sea the availability of ringed seals to polar bears. To hunt ice. Large adult males also declined in weight during successfully, polar bears must locate ringed seals in the 1980s (Derocher 1991). This does not support or habitat favorable for catching them. The distribution of necessarily contradict the exclusion hypothesis. seals under the winter ice is not known for Hudson In many large mammals, density-dependent changes Bay. Similarly, the affect of ice types on the hunting occur at population levels close to the carrying capacity success of polar bears has not been investigated. The (Fowler 1981a,b). The hypothesis that the density of onset of freeze-up, break-up, coverage, and ice polar bears suddenly became sufficient to trigger morphology in space and time are quite variable for density effects in the early 1980s, almost 20 years after arctic sea ice. The long life span and life-history the introduction of harvest controls would seem initially characteristics of both polar bears and ringed seals to have merit. However, population numbers appeared mitigate the effects of environmental variability on the to be constant over the period (1980-90) when mean density of both populations, although the effects on weight, recruitment rate, and cub survival rate declined production of cubs and survival of cubs can be severe (Derocher 1991). Additionally, this population has (Stirling et al. 1977a). The mean temperature in been harvested at approximately the maximum sustained eastern Hudson Bay was colder than normal during the yield for the past 15 years (Taylor et al. 1987a, Lee 1980's (Findlay and Deptuch-Stapf 1991). The and Taylor 1994). The hypothesis that density unusually low temperatures could have reduced seal increased, causing a decline in cub survival and productivity or availability. recruitment is, therefore, not supported by the existing Ifthe carrying capacity of polar bears in Hudson Bay data. However, if the onset of density effects occurred has been reduced by reductions in either seal numbers very rapidly at some threshold, and the change in or seal availability, the declines in cub survival and density required to trigger density effects was quite recruitment could be regarded as a density-dependent small, it is possible that a sufficient increase in density DENSITY EFFECTS ON BEARS • Taylor et at. 27 did occur but was not detected (Derocher 1991). bear cub survival in the Svalbard area. The available Infanticide, exclusion from prime habitats, a long­ data were insufficient to yield any insight on the term density-independent fluctuation in seal availability biological mechanisms causing low cub survival in this or abundance, a reduction in the Hudson Bay carrying area. It is particularly unfortunate that there is no capacity for polar bears, and increased density of polar density data over the suggested recovery period for this bears are all possible explanations for the declines in population. maternal weight, cub survival, and recruitment observed. COMPARISONS AMONG POPULATIONS Eastern Beaufort Sea Most population inventory work has been done as In contrast to the Churchill population, information part. of a management mandate on harvested on both polar bears and seals was available from long­ populations. Because these populations have been term studies (1970-87) in the Beaufort Sea. The reduced by harvest, the densities may not be sufficient numbers and productivity of ringed seals were reduced to cause density effects. The polar bear harvest in most in 1974 and 1975 (Stirling et al. 1977a). The decline areas of North America is selective for males (Lee and in seals resulted in a reduction in the body weights of Taylor 1994), which could reduce the likelihood of polar bears and the number of female polar bears that infanticide and intraspecific predation. The mechanism were accompanied by offspring (Stirling et al. 1976, by which density acts to regulate polar bear numbers is 1977a; Kingsley 1979; Amstrup et al. 1986). The age still a matter of conjecture. To assess the effects of distributions after 1975 indicated a significant density on population dynamics adequately, the full (P < 0.05) reduction in the cohorts born in 1974 and compliment of population parameters should be 1975, but not in the annual rate of survival of adult measured under varying densities. However, densities polar bears (Stirling et al. 1988, unpubI. data). The of most polar bear populations have been determined reproductive rates of polar bears returned to previous for periods that were too short to detect changes in levels when ice conditions apparently improved in abundance, recruitment, or survival rates. subsequent years (Stirling et al. 1988), suggesting that One possibility for examining the existing data for the observed reduction in reproductive rate was not due correlations which might suggest biological mechanisms to a long-term reduction in seal numbers. Although for density effects is to compare population Amstrup et aI. (1986) suggest that population-growth characteristics from populations at different densities. rate was diminished for the heavy ice years, and As previously mentioned in both the black and brown indicated that polar bear numbers were reduced, the bear sections, this type of analysis can be difficult to population did not decline perceptibly (Stirling et al. interpret. Prey density, weather, ice conditions, 1988). The Beaufort Sea example indicates the oceanographic characteristics, and harvest differs vulnerability of recruitment and cub survival rates to among study areas, and may affect population vital environmental perturbation, and suggests that these rates independent of bear density. Additionally, parameters can be reduced in one or several years methodologies and the potential for sampling bias without a concurrent reduction in adult survival or a differ among studies. As a result, differences in large (> 10 %) decrease in population number. population characteristics may be due to factors other than density. Svalbard These difficulties were partially mitigated by using a The Svalbard population has not been hunted since single analysis procedure (Taylor et al. 1987b) on all 1972 (Larsen 1985), and may have increased to a data. Densities of polar bears on the spring sea ice number large enough to be regulated by density effects. were determined by dividing the population estimates by Although data for this population are limited and the habitat area which was defined as the ice area statistical comparisons are not possible, annual cub between the minimum ice cover in summer and the survival has been estimated at 0.44 (calculated from shoreline/maximum winter ice cover (Taylor and Lee Larsen 1985) which is dramatically lower than 0.77, the 1994). The mean litter size, individual cub mortality average cub survival rate from data pooled from across rate (loss of cubs as individuals), litter cub survival rate Canada (Taylor et al. 1987a) and for Churchill (loss of cubs as whole litter mortality), litter size, (Derocher 1991). Larsen (1985) suggested that average age of first reproduction, and Chapman-Robson unfavorable ice conditions or cannibalism by adult age distribution survival rate for adults (age > 4) (4)) males might be responsible for the low rate of polar were calculated according to Taylor et al. (l987b) when 28 Int. Con! Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp. data were sufficient. The inter-litter interval was not have been reduced by harvest to levels where density used because direct measures of this parameter are effects would be so minor they would not be detectable. almost exclusively restricted to the Churchill population No differences in habitat quality (i.e., carrying for polar bears. We looked for correlations between capacity) were considered in these comparisons. these survival and reproduction estimates and population Although these continental comparisons do not suggest density that might indicate density effects. density effects for polar bear densities ranging from 1 None of the parameter estimates was significantly to 7 bears per 1,000 km2, these data are insufficient to (P < 0.05) correlated with density estimates (Fig. 6, demonstrate the lack of density effects in the polar bear Table 6), possibly because all the populations included populations considered.

1.00~------..., 2.00....------. El • • HI ~ 1.90 0.95 •Al Al en 1.80 02 0.90 A:/" a:: • w • 1.70 01 El• •HI ~ 0.85 •F 01• 1.60 •02 ::::i

0.80'---...... ----...... - ...... 1.501...---''--...... 0...-...... ----1

0.90~------. 1.00r------, •F 0.80 0.95 02 • •Al • 0.70 s: 0.90 A2 •01 F. •Al 0.60 0.85

0.501...--...... - ...... - ...... - ..... 0.80'---...... - ...... - ...... -.....l

1.00r------, 8.00r------,

F w· •Al e" 0 7.00 •

0.40 '---...... ----...... - ...... 4.001...---''----''----''-----1 0.0 2.0 4.0 6.0 8.0 0.0 2.0 4.0 6.0 8.0 DENSITY DENSITY 2 (bears per 1000 km ) (bears per 1000 km 2)

Fig. 6. Individual cub survival. whole litter cub survival. total cub survival. litter size. mean age of first reproduction. and adult survival rate (as Chapman -Robson <1>1 were compared to polar bear density on the sea ice for 8 of North Americas's 12 polar bear populations. The populations are identified by management area designations (Canadian Federal-Provincial Polar Bear Technical Committee): A 1 (Derocher 1991). A2 (Kolenosky pers. commun.). 01 and E1 (unpublished data for NE Baffin and Viscount Melville Sound). 02 (Stirling et al. 1980). F (Schweinsburg et al. 1981). H1 (Amstrup et al. 1986). H2 (Stirling et al. 1988). The correlation coefficient (T) and observed level of significance (P) are listed for each parameter type in Table 6. DENSITY EFFECTS ON BEARS • Taylor et al. 29

Table 6. Individual cub mortality rate (loss of cubs as individualsl. litter cub mortality rate (loss of cubs as whole litter mortality), total cub survival (cub survival rate x whole litter survival ratel. litter size, mean age of first reproduction, and adult survival rate (as Chapman -Robson c&» were compared with polar bear density for 8 of North America's 12 polar bear management units that had sufficient data. The data are identified by management area designations (Canadian Federal-Provincial Polar Bear Technical Committee). The correlation coefficient (f) and observed level of significance (P) are listed for each parameter type.

Individual Litter Total Mean age Area Density Ca) cub survival cub survival cub survival Litter size 1st repro. Alb 3.2 0.911 0.627 0.571 1.89 0.923 A2c 3.0 0.909 0.535 0.486 1.58 0.909 Did 1.0 0.847 1.73 0.894 5.6 D2e 2.3 0.838 1.76 0.932 5.2 El d 5.0 0.980 0.620 0.608 1.72

Ff 5.9 0.905 0.834 0.878 1.64 0.879 7.2 Hl g 7.0 0.970 1.72 0.888 6.5 H2h 6.5 0.830 0.741 0.615 1.81 0.849 7.0 Pearson r 0.428 0.821 0.638 -0.031 -0.673 0.633 p 0.290 0.088 0.247 0.942 0.097 0.177 a Bears per 1,000 km'. b Churchill, Western Hudson Bay (Derocher 1991). c Eastern Hudson Bay (G.B. Kolenosky, Ont. Minist. Nat. Resour., pers. commun). d Northeastern Baffin (Dl) or Viscount Melville Sound (E2) (unpublished data). e Southeastern Baffin (Stirling et al. 1980). f Lancaster Sound (Schweinsburg et a!. 1981). g Southwestern Beaufort Sea (Amstrup et al. 1986). h Northeastern Beaufort Sea (Stirling et al. 1988). Cells left blank (-) did not have accurate estimates because of non-representative sampling or insufficient data.

SUMMARY role of nutritional stress during periods without access Our review was unable to identify any popUlation of to food in population regulation cannot be assessed. polar bears in which density effects were demonstrated. Animals have physiological, metabolic, morphological, We suggest that cub survival and recruitment are the and behavioral mechanisms by which they can population parameters most sensitive to environmental compensate for periods without food and the intensity fluctuation, and would therefore be the most likely of nutritional stress has often been exaggerated (King mechanisms for nutritional density effects. In large and Murphy 1985). mammals, malnutrition is often a major mortality factor Several authors have speculated that adult polar bear during the first year of life (Sinclair 1977, Guinness et males might influence cub survival through infanticide, al. 1978, Packer et al. 1988). The influence of food intraspecific predation, and exclusion from prime supply may also be critical in polar bear populations. feeding areas (Lunn and Stenhouse 1985, Taylor et al. The condition of cubs strongly influences their survival 1985, Larsen 1985). However, no study has which is strongly correlated with the condition of the demonstrated that infanticide or intraspecific predation mother (Derocher 1991) and ultimately, related to the is a significant cause of cub or adult mortality, or that availability of ringed seals (Stirling et al. 1976, 1977a; rates of infanticide or intraspecific predation increase Kingsley 1979; Amstrup et al. 1986). During the 2 with increased density. major periods that polar bears undergo extended fasts, A tendency for females, particularly females with over-winter and during the autumn, maternal fat cubs, to use different areas than males has been noted reserves are crucial for cub survival. (Derocher and Stirling 1990, Stirling et al. 1993). No No polar bear study to date has shown a relationship study has offered any evidence that the spatial between density and female condition. Until a better segregation observed resulted in reduced weights or understanding of polar bear energetics is available, the increased mortality rates. Exclusion from prime 30 Int. Cont Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp. habitats is an unsubstantiated although viable hypothesis evaluate this hypothesis, and to determine if the for density-dependent population regulation in polar mechanism is mainly nutritional, mainly social, or bears. Similarly, the degree to which subadult polar socially mediated habitat partitioning with nutritional bears rely on carrion is unknown. There is evidence of consequences. Studies in Svalbard, Siberia, and subadult bears being displaced from both carrion and Novaya Zemlya, where polar bears are not harvested, kills by older, larger bears (Stirling 1974). Increased could provide more insight into regulating mechanisms. densities may intensify competition for seal kills and All North American and Greenland polar bear carrion, causing reduced subadult survival and possibly populations are harvested and, with the possible delayed reproductive maturity. exception of the western Hudson Bay (i.e., Churchill) We hypothesize that density-dependent cub (and population, have been reduced to sufficiently low possibly subadult) mortality and diminished numbers that density effects are probably negligible reproduction are the factors which regulate polar (Appendix I). bear populations. The data are presently insufficient to DENSITY EFFECTS ON BEARS • Taylor et al. 31

LITERATURE CITED Manage. 4:201-204. ALT. G.L. 1978. Dispersal patterns of black bears in ___. 1983. Population characteristics of black bears in northeastern Pennsylvania-a preliminary report. East. west central Idaho. J. Wild!. Manage. 47:405-412. Workshop Black Bear Res. and Manage. 4:186-199. BEEMAN, L.E., AND M.R. PELTON. 1980. Seasonal foods ___. 1980. Rate of growth and size of Pennsylvania and feeding ecology of black bears in the Smoky black bears. Penn. Game News 51(12):7-17. Mountains. Int. Conf. Bear Res. and Manage. 4: 141-147. 1981. Reproductive biology of black bears of BLAFFER-HRDY, S. 1977. The langurs of Abu: female and northeastern Pennsylvania. Trans. N.E. Fish and Wild!. male strategies of reproduction. Harvard Univ. Press. Conf. 38:88-89. Cambridge Mass. 361pp. ___. 1984. Black bear cub mortality due to tlooding of BLANCHARD, B. M. 1987. Size and growth patterns of the natal dens. J. Wild!. Manage. 48:1432-1434. Yellowstone grizzly bear. Int. Conf. Bear Res. Manage. ___. 1989. Reproductive biology of female black bears 7:99-107. and early growth and development of cubs in northeastern BODENHIEMER, F.S. 1938. Problems of animal ecology, Pennsylvania. Ph.D. Thesis, W.Va. Univ., Morgantown. Oxford Univ. Press, London. 116pp. BUNNELL, F.L., AND D.E.N. TAIT. 1980. Bears in models , AND J. M. GRUTTADAURIA. 1984. Reuse of black and in reality - implications to management. Int. Conf. bear dens in northeastern Pennsylvania. J. Wildl. Manage. Bear Res. and Manage. 4:15-23. 48 :236-239. __' AND __. 1981. Population dynamics of bears ­ AMSTRUP, S., 1. STIRLING, ANDJ.W. LENTFER. 1986. Past implications. Pages 75-98 in C.W. Fowler and T.D. and present status of polar bears in Alaska. Wild!. Soc. Smith, eds. Dynamics oflarge mammal populations. John Bull. 14:241-254. Wiley and Sons, New York. ANDERSON, R.K. 1989. A proposal to study the population CALIFORNIA DEPARTMENT OF FISH AND GAME. 1991. Draft dynamics, behavior, and habitat of black bears on Stockton environmental document regarding bear hunting. Calif. Island, Apostle Islands National Lakeshore. Unpub!. rep., Dep. Fish and Game, Sacramento. 233pp + 8 append. Univ. Wis., Stevens Point. 15pp. CARNEY, D.W. 1985. Population dynamics and denning , AND D. TRAUBA. 1991. Black bear density on ecology of black bears in Shenandoah National Park, Stockton Island and mainland Wisconsin. Unpub!. rep., Virginia. M.S. Thesis, Va. Polytech. Inst. and State Univ. Wis., Stevens Point. 9pp. Univ., Blacksburg. 84pp. ANDREWARTHA, H.G., AND L.C. BIRCH. 1954. The CLARK, J.D. 1991. Ecology of two black bear (Ursus distribution and abundance of animals. Univ. of Chicago americanus) populations in the interior highlands of Press, II!. Arkansas. Ph.D. Thesis, Univ. Ark, Fayetteville. 228pp. AUNE, K., AND B. BRANNON. 1987. East Front grizzly COY, P.L., AND D.L. GARSHELIS. 1992. Reconstructing studies. Mont. Dep. Fish, Wild!. Parks, Helena. 195pp. reproductive histories of black bears __' AND W. KASWORM. 1989. Final Report. East Front from the incremental layering in dental cementum. Can. Grizzly Studies. Mont. Dep. Fish, Wild!. Parks. Helena. J. Zoo!. 70:2150-2160. 332pp. CRAIGHEAD, J.1., F.C. CRAIGHEAD JR., AND J. SUMNER. __' M. MADEL, AND C. HUNT. 1986. Rocky Mountain 1976. Reproductive cycles and rates front grizzly bear monitoring and investigation. Mont. in the grizzly bear Ursus arctos horribilis, of the Dep. Fish, Wild!. Parks, Helena. 175pp Yellowstone Ecosystem. Int. Conf. Bear Res. Manage. BARBER, K.R., AND F.G. LiNDZEY. 1986. Breeding 3:337-356. behavior of black bears. Int. Conf. Bear Res. and , K. GREER, R.R. KNIGHT, AND H. PAC. 1988. Manage. 6: 129-136. Grizzly bear mortalities in the Yellowstone Ecosystem. BARTMANN, R.M., G.e. WHITE, AND L.H. CARPENTER. Wildlands Institute, Missoula, MT. 118pp. 1992. Compensatory mortality in a Colorado mule deer , J.R. VARNEY, AND F.e. CRAIGHEAD JR. 1974. A population. Wild!. Monogr. 121. 39pp. population analysis of the Yellowstone grizzly bears. BECK, T.D.1. 1991. Black bears of west-central Colorado. Mont. For. and Cons. Exp. Stn. Bull. 40. Univ. Montana, Colo. Div. Wild!. Tech. Pub!. 39. 86pp. Missoula. 20pp. BEECHAM, J.J. 1980a. Population characteristics, denning, DARWIN, C. 1859. The origin of species by means of and growth patterns of black bears in Idaho. Ph.D. natural selection. Reprinted by The Modern Library, Thesis, Univ. Montana, Missoula. 101pp. Random House, New York, N.Y. 198Gb. Some population characteristics of two DEAN, F.e., L.M. DARLING, AND A.G. LiERHAUS. 1986. black bear populations in Idaho. Int. Conf. Bear Res. and Observations of intraspecific killing by brown bears, Ursus 32 Int. Con! Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

arctos. Can. Field. Nat. 100:208-211. 1987. A review of density dependence in DEN BOER, P.I. 1987. The density dependence debate: populations of large mammals. Pages 401-441 in H. H. detecting density dependence. Trends Ecol. Evol. 2:77-78. Genoways, ed. Current Mammalogy, vol. 1. Plenum DEROCHER, A.E. 1991. Population dynamics and ecology of Press, New York, N.Y. polar bears in western Hudson Bay. Ph.D. Thesis, Univ. 1990. Density dependence in northern fur seals Alberta, Edmonton, Can. 189pp. (Callorhinus ursinus). Marine Mamm. Sci. 6:171-195. ___' D. ANDRIASHEK, AND I.P.Y. ARNOULD. 1993. GARSHELIS, D.L. 1989. Nuisance bear activity and Aspects of milk composition and lactation in polar bears. management in Minnesota. Pages 169-180 in M. Bromley, Can. I. Zool. 71:561-567. ed. Bear people conflicts-proceedings of a symposium on ___' R.A. NELSON, I. STIRLING, AND M.A. RAMSAY. management strategies. Northwest Territ. Dep. Renew. 1990. Effects of fasting and feeding on serum urea and Resour., Yellowknife. serum creatinine levels in polar bears. Marine Mamm. __'_. 1991. Monitoring effects of harvest on black bear Sci. 6: 196-203. populations in North America: a review and evaluation of ---,AND I. STIRLING. 1990. Distribution of polar bears techniques. East. Workshop Black Bear Res. and Manage. (Ursus maritimus) during the ice-free period in western 10:120-144. Hudson Bay. Can. I. Zool. 68:1395-1403. 1992. Mark-recapture density estimation for ___' and . 1992. The population dynamics of animals with large home ranges. Pages 1,098-1,111 in polar bears in western Hudson Bay. In D.R. McCullough D.R. McCullough and R. H. Barrett, eds. Wildlife 2001: and R.H. Barrett, eds. Wildlife 2001: Populations. Populations. Elsevier Applied Science, London. Elsevier Science Publishers, London. __. 1993. Monitoring black bear populations: pitfalls DODD, A.E., R.D. BRANNON, AND R.D. MACE. 1986. Final and recommendations. West. Black Bear Workshop 4: 123­ programmatic environmental impact statement: the grizzly 144. bear in northwestern Montana. Mont. Dep. Fish, Wildl. __' K.V. NOYCE, AND P.L. COY. 1988. Ecology and and Parks, Helena. 287pp. population dynamics of black bears in north-central DUNBAR, M.I. 1981. Physical causes and biological Minnesota. Pages 43-53 in B. Joselyn, ed. Summaries of significance of polynyas and other open water in sea ice. wildlife research project findings, 1987. Minn. Dep. Nat. In I. Stirling and H. Cleator, eds. Polynyas in the Resour., St. Paul. Canadian Arctic. Can. Wildl. Servo Occas. Pap. 45:29-40. ,AND M.R. PELTON. 1981. Movements of black bears EBERHARDT, L.L., AND D.B. SINIFF. 1977. Population in the Great Smoky Mountains National Park. J. Wildl. dynamics and marine mammal management. J. Fish. Res. Manage. 45:912-925. Board of Can. 34:183-190. GAUSE, G.J. 1934. The struggle for existence. Williams and EGBERT, A.L., AND A.W. STOKES. 1976. The social Wilkins, Baltimore, Md. behavior of brown bears on an Alaska salmon stream. Int. GILL, R.B., AND T.D.I. BECK. 1991. Bears: analysis of Conf. Bear Res. Manage. 3:41-56. season structure alternatives. Colo. Div. Wildl., unpubl. EILER, I.H., W.G. WATHEN, AND M.R. PELTON. 1989. rep., Fort Collins. 31pp. Reproduction in black bears in the southern Appalachian GINZBURG, L.R., S. FERSON, AND H.R. AKCAKAYA. 1990. mountains. I. Wildl. Manage. 53:353-360. Reconstructibility of density dependence and the ELOWE, K.D., AND W.E. DODGE. 1989. Factors affecting conservative assessment ofextinction risks. Conserv. BioI. black bear reproductive success and cub survival. J. 4:63-70. Wildl. Manage. 53:962-968. GLENN, L.P., I.W. LENTFER, I.B. FARO, AND L.H. MILLER. ERICKSON, A.W. 1957. Techniques for live-trapping and 1976. Reproductive biology of female brown bears, handling black bears. Trans. North Am. Wildl. Conf. (Ursus arctos) , McNeil River, Alaska. Int. Conf. Bear 22:520-543. Res. Manage. 3:381-390. FINDLAY, B.F., ANDA. DEPTUCH-STAPF. 1991. Colder than GRABER, D.M. 1981. Ecology and management of black normal temperatures over northeastern Canada during the bears in Yosemite National Park. Ph.D. Thesis, Univ. 1980s. Climatic Persp. 13:9-12. Calif., Berkeley. 206pp. FOWLER, C.W. 1981a. Comparative population dynamics in GUINNESS, F.E., T.H. CLUTTON-BROCK, AND S.D. ALBON. large mammals. Pages 437-455 in C. W. Fowler and T. 1978. Factors affecting calf mortality in red deer (Cervus D. Smith, eds. Dynamics of large mammal populations. elaphus). I. Anim. Ecol. 47:817-832. John Wiley and Sons, New York, N.Y. HAMMILL, M.D., AND T.G. SMITH. 1991. The role of ___. 1981b. Density dependence as related to life history predation on the ecology of ringed seals in Barrow Strait, strategy. Ecology 62:602-610. N.W.T., Canada. Marine Mamm. Sci. 7:123-135 DENSITY EFFECTS ON BEARS • Taylor et al. 33

HASSELL, M.P. 1986. Detecting density dependence. KINGSLEY, M.C.S. 1979. Fitting the von Bertalanffy growth Trends Eco\. Evo\. 1:90-93. equation to polar bear age-weight data. Can. J. Zoo\. __' AND M.W. SABELIS. 1987. The density dependence 60:40-44. debate: reply from M.P. Hassell and M.W. Sabelis. KLENNER, W. 1987. Seasonal movements and home range Trends Eco\. Evo\. 2:78. utilization patterns of the black bear, Ursus americanus, in HAUSFATER, G., AND S. BLAFFER-HRDY, EDS. 1985. western Manitoba. Can. Field-Nat. 101 :558-568. Infanticide: comparative and evolutionary perspectives. KNIGHT, R.R., B.M. BLANCHARD, AND L.L. EBERHARDT. Aldine, New York, N.Y. 598pp. 1988a. Mortality patterns and population sinks for HELLGREN, E.C., AND M.R. VAUGHAN. 1989. Yellowstone grizzly bears, 1973-1985. Wild\. Soc. Bull. Demographic analysis of a black bear population in the 16:121-125. Great Dismal Swamp. 1. Wild\. Manage. 53:969-977. __' __' AND D.1. MATTSON. 1988b. Yellowstone HERRERO, S. 1983. Social behavior of black bears at a grizzly bear investigations. Rep. of the Interagency Study garbage dump in Jasper National Park. Int. Conf. Bear Team. 1987. U.S. Nat. Park Serv., Bozeman. Mont. Res. and Manage. 5:54-70. 80pp. HESSING, P., AND L. AUMILLER. 1994. Observations of __, __' AND __. 1989. Yellowstone grizzly bear conspecific predation by brown bears, Ursus arctos, in investigations. Rep. of the Interagency Study Team. Alaska. Can. 1. Zoo\.: In press. 1988. U.S. Nat. Park Serv., Bozeman. Mont. 34pp. HESTBECK, J.B. 1982. Population regulation of cyclic __' __' AND . 1990. Yellowstone grizzly bear mammals: the social fence hypothesis. Oikos, 39:157-163. investigations. Rep. of the Interagency Study Team. HOWARD, L.O., AND W.F. FISKE. 1911. The importation 1987. U.S. Nat. Park Serv., Bozeman, Mont. 33pp. into the United States of the parasites of the gypsy moth __' AND L.L. EBERHARDT. 1985. Population Dynamics and the brown-tailed moth. U.S. Dep. Agric. Bur. Ent. of Yellowstone Grizzly Bears. Ecology 66:323-334. Bull. 91:1-312. KOLENOSKY, G.B. 1986. The effects of hunting on an HUGIE, R.D. 1982. Black bear ecology and management in Ontario black bear population. Int. Conf. Bear Res. and the northern conifer-deciduous forests of Maine. Ph.D. Manage. 6:45-55. Thesis, Univ. Mont., Missoula. 203pp. __. 1990. Reproductive biology of black bears in east­ JONKEL, C.J. 1967. Life history, ecology, and biology of central Ontario. Int. Conf. Bear Res. and Manage. 8:385­ the polar bear, autumn 1966 studies. Can Wildl. Servo 392. Prog. Note No.1. 8pp. KREBS, C.J. 1978. Ecology: the experimental analysis of __, AND I.M. COWAN. 1971. The black bear in the distribution and abundance. Second ed. Harper & Row spruce-fir forest. Wildl. Monogr. 27. 57pp. PubI. , New York, N.Y. 678pp. KAUTZ, J.E. 1990. Testing for compensatory responses to , M.S. GAINES, B.L. KELLER, J.H. MYERS, AND R.H. removals from wildlife populations. Trans. North Am. TAMARIN. 1973. Population cycles in small rodents. Wild\. Nat. Resour. Conf. 55:527-533. Science 179:35-41. KEATING, K.A. 1986. Historic grizzly bear trends in Glacier LACK, D. 1967. The natural regulation of animal numbers. National Park. Montana Wildl. Soc. Bull. 14:83-87. Clarendon Press, Oxford, Eng. 343pp. KEAY, J.A. 1990. Black bear population dynamics in LARSEN, T. 1985. Abundance, range and population biology Yosemite National Park. Coop. Nat. Park Resour. Studies of the polar bear (Ursus maritimus) in the Svalbard area. Unit, Univ. Calif., Davis, Tech. Rep. No. 39. 139pp. Ph.D. Thesis, Norsk Polarinstitutt, Oslo. Norway. 247pp. KEMP, G.A. 1972. Black bear population dynamics at Cold LATOUR, P.B. 1981. Spatial relationships and behavior of Lake, Alberta, 1968-1970. Int. Conf. Bear Res. and polar bears (Ursus maritimus Phipps) concentrated on land Manage. 2:26-31. during the ice-free season of Hudson Bay. Can. 1. Zoo\. __. 1976. The dynamics and regulation of black bear 59:1763-1774. Ursus americanus populations in northern Alberta. Int. LECOUNT, A.L. 1982. Characteristics of a central Arizona Conf. Bear Res. and Manage. 3: 191-197 . black bear population. 1. Wild\. Manage. 46:861-868. KENDALL, K.C. 1985. Grizzly bear population trend __. 1984. Black bear cub production and survival in studies, Apgar Mountains, Glacier National Park. Nat. central Arizona. Fed. Aid Wildl. Restor. Proj. W-78-R Park Serv. Prog. Rep. for 1984. Glacier National Park, Final Rep. Arizona Game and Fish Dep. lOpp. Mont. 23pp. __. 1987a. Causes of black bear cub mortality. Int. KING, I.R., AND M.E. MURPHY. 1985. Periods of Conf. Bear Res. and Manage. 7:75-82. nutritional stress in the annual cycles of endotherms: fact . 1987b. Characteristics of a northern Arizona black or fiction? Am. Zoo\. 25:955-964. bear population. Fed. Aid Wildl. Restor. Proj. W-76-R 34 Int. Con! Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

Final Rep. 26pp. T.D. Smith, eds. Dynamics oflarge mammal populations. 1990. Characteristics of an east-central Arizona John Wiley and Sons, New York, N.Y. black bear population. Arizona Game and Fish Dep. __. 1986. The Craigheads' data on Yellowstone grizzly Tech. Rep. No.2. 32pp. bear populations: relevance to current research and LEE, J.L., AND M.K. TAYLOR. 1994. Aspects of the polar management. Int. Conf. Bear Res. and Manage. 6:21-32. bear harvest in the Northwest Territories, Canada. Int. McLAUGHLIN, C.R., G.1. MATULA, JR., AND R.1. Conf. Bear Res. and Manage. 9:In Press. O'CONNOR. 1994. Synchronous reproduction by Maine LEFRANC, JR., M.N., M.B. Moss, K.A. PATNODE, AND black bears. Int. Conf. Bear Res. and Manage. 9:In Press. W.C. SUGG, EDS. 1987. Grizzly Bear Compendium. McLEAN, P.K. 1991. The demographic and morphological Interagency Grizzly Bear Comm., Washington, D.C. characteristics of black bears in the Smoky Mountains. 540pp. Ph. D. Thesis, Univ. Tennessee, Knoxville. 156pp. LiNDZEY, F.G., K.R. BARBER, R.D. PETERS, AND E.C. __, AND M.R. PELTON. 1994. Estimates of population MESLOW. 1986. Responses of a black bear population to density and growth of black bears in the Smoky a changing environment. Int. Conf. Bear Res. and Mountains. Int. Conf. Bear Res. and Manage. 9:In Press. Manage. 6:57-63. McLELLAN, B.N. 1989a. Population dynamics of grizzly , AND E.C. MESLOW. 1977a. Population bears during a period of resource extraction development. characteristics of black bears on an island in Washington. II. Morality rates and causes. Can. J. Zoo!. J. Wild!. Manage. 41:408-412. 67:1861-1864. __' AND __. 1977b Home range and habitat use by · 1989b. Population dynamics of grizzly bears during black bears in southwestern Washington. 1. Wild!. a period of resource extraction development. III. Natality Manage. 41 :413-425. and rate of change. Can. J. Zoo!' 67:1865-1868. LINDZEY, J.S., G.L. ALT, C.R. McLAUGHLIN, AND W.S. __. 1989c. Population dynamics of grizzly bears during KORDEK. 1983. Population response of Pennsylvania a period of resource extraction development. I. Density black bears to hunting. Int. Conf. Bear Res. and Manage. and age/sex structure. Can. J. Zoo!. 67:1856-1860. 5:34-39. MESLOW, E.C., AND L.B. KEITH. 1968. Demographic LOMNICKI, A. 1987. The density dependence debate: parameters of a snowshoe hare population. J. Wild!. density-vague ecology. Trends Eco!. Evo!. 2:76. Manage. 32:812-834. LUNN, N.J., AND G.B. STENHOUSE. 1985. An observation MESSIER, F. 1991. Detection of density-dependent effects on of possible cannibalism by polar bears (Ursus maritimus). caribou numbers from a series of census data. Rangifer. Can. J. Zoo!. 63:1516-1517. 11 :36-46. , AND I. STIRLING. 1985. The significance of MILLER, S.D. 1987. Big game studies: Vo!. VI, Black bear supplemental food to polar bears during the ice-free period and brown bear. Susitna hydroelectric project final report. of Hudson Bay. Can. 1. Zoo!. 63:2291-2297. Alaska Dep. Fish and Game, and Alaska Power Authority. MALTHUS, R.T. 1798. An essay on the principle of 276pp. population as it affects the future improvement of society. · 1990a. Detection of differences in brown bear Johnson, London. density and population composition caused by hunting. MARTINKA, C.J. 1974. Population characteristics of grizzly Int. Conf. Bear Res. Manage. 8:393-404. bears in Glacier National Park, Montana. 1. Mammal. 1990b. Impact of increased bear hunting on 55:21-29. survivorship of young bears. Wild!. Soc. Bul!. MATTSON, D.J. 1990. Human impacts on bear habitat use. 18 :462-467. Int. Conf. Bear Res. and Manage. 8:33-56. · 1990c. Impact of increased hunting pressure on the , B.M. BLANCHARD, AND R.R. KNIGHT. 1991. Food density, structure, and dynamics of brown bear populations habits of Yellowstone grizzly bears, 1977-87. Can. 1. in Alaska's Game Management Unit 13. Proj. W-23-3, Zoo!' 69:1619-1629. Study 4.21. Alaska Dep. Fish and Game, Juneau. 88pp. __' R.R. KNIGHT, AND B.M. BLANCHARD. 1987. The __. 1990d. Population management of bears in North effects of developments and primary roads on grizzly bear America. Int. Conf. Bear Res. and Manage. 8:357-373. habitat use in Yellowstone National Park, Wyoming. Int. · 1994. Black bear reproduction and cub survivorship Conf. Bear Res. Manage. 7:259-273. in south-central Alaska. Int. Conf. Bear Res. and Manage. , AND M.M. REID. 1991. Conservation of the 9:In Press. Yellowstone grizzly bear. Cons. Bio!. 5:363-372. , E. F. BECKER, AND W. G. BALLARD. 1987. Black MCCULLOUGH, D.R. 1981. Population dynamics of the and brown bear density estimates using modified capture­ Yellowstone grizzly. Pages 173-196 in C.W. Fowler, and recapture techniques in Alaska. Int. Conf. Bear Res. and DENSITY EFFECTS ON BEARS • Taylor et al. 35

Manage. 7:23-35. PEARSON, A.M. 1975. The northern interior grizzly bear. __, AND R.A. SELLERS. 1990. Brown bear density on the Can. Wildl. Servo Rep. Ser. No. 34. 86pp. Alaskan Peninsula at Black Lake, Alaska. Alaska Dep. PELCHAT, B.a., AND R.L. RUFF. 1986. Habitat and spatial Fish and Game, Juneau. 30pp. relationships of black bears in boreal mixedwood forest of MILLER, S.J., N. BARICHELLO, AND D. TAIT. 1982. The Alberta. Int. Conf. Bear Res. and Manage. 6:81-92. grizzly bears of the MacKenzie Mountains, Northwest PELTON, M.R. 1989. The impacts of oak mast on black Territories. Completion Rep. 3. 118pp. bears in the southern Appalachians. Pages 7-11 in C.E. MURDOCK, W.W. 1970. Population regulation and McGee, ed. Proc. Workshop Southern Appalachian mast population inertia. Ecology 51 :497-502. management, U.S. Dep. of Agric. For. Servo and Univ. MURIE, A. 1981. The grizzlies of Mount McKinley. Tenn., Knoxville. U.S.D.1. Nat. Park Servo Sci. Monogr. Ser. No. 14. PICTON, H.D. 1978. Climate and the reproduction of grizzly 251pp. bears in Yellowstone National Park. Nature (London) MURRAY, B.G., JR. 1982. On the meaning of density 274:888-889. dependence. Oecologia 53:370-373. __' AND R.R. KNIGHT. 1986. Using climate data to NAGY, J.A., ANDM.A. HAROLDSON. 1990. Comparisons of predict grizzly bear liller size. Int. Conf. Bear Res. some home range and population parameters among four Manage. 6:41-44. grizzly bear populations in Canada. Int. Conf. Bear Res. POLLARD, E., K.H. LAKHANI, ANDP. ROTHERY. 1987. The Manage. 8:227-235. detection of density dependence from a series of annual __., R.H. RUSSEL, A.M. PEARSON, M.C. KiNGSLEY, AND censuses. Ecology 68:2046-2055. B. C. GOSKI. 1983a. Ecological studies of the grizzly bear POLLOCK, K.H., J.D. NICHOLS, C. BROWNIE, AND J.E. in the arctic mountains, northern Yukon Territory, 1972 to HINES. 1990. Statistical inference for capture-recapture 1975. Can. Wildl. Ser. Rep. 104pp. experiments. WildI. Monogr. 107. 97pp. __' __, __, __, AND C.B. LARSEN. 1983b. A RAMSAY, M.A., AND I. STIRLING. 1986. On the mating study of grizzly bears on the barren grounds of system of polar bears. Can. J. ZooI. 64:2142-2151. Tuktoyaktuk Peninsula and Richards Island, Northwest __, AND __. 1988. Reproductive biology and ecology Territories, 1974 to 1978. Can. Wildl. Ser. Rep. 136pp. of female polar bears (Ursus maritimus). J. Zool., London NICHOLSON, AJ. 1954a. Compensatory reactions of 214:601-634. populations to stress, and their evolutionary significance. REYNOLDS, D.G., AND J.J. BEECHAM. 1980. Home range Aust. J. ZooI. 2:1-8. activities and reproduction of black bears in west-central 1954b. An outline of the dynamics of animal Idaho. Int. Conf. Bear Res. and Manage. 4:181-190. populations. Aust. J. ZooI. 2:9-65. REYNOLDS, H.V. 1976. North Slope grizzly bear studies. NOYCE, K.V., AND D.L. GARSHELIS. 1994. Body size and Final Report, Project W-17-6 and W-17-7, lobs 4.8R, blood characteristics as indicators of condition and 4.1R, and 4.11R. 20pp. reproductive performance in black bears. Int. Conf. Bear __. 1989. Population dynamics of a hunted grizzly bear Res. and Manage. 9:In Press. population in the north-central Alaska Range. Project OLSON, T.L. 1993. Infanticide in brown bears, Ursus W-23-1. Job 4.19, Alaska Dep. Fish and Game, Juneau. arctos, at Brooks River, Alaska. Can. Field. Nat. 107:92­ 63pp. 94. __' ANDT.A. BOUDREAU. 1990. Effects of harvest rates PACKER, C., L. HERBST, A.E. PUSEY, J.D. BYGOTT, J.P. on grizzly bear population dynamics in the north-central HANBY, S.J. CAIRNS, AND H.B. MULDER. 1988. range. Project W-23-3, Study 4.19. Alaska Dep. Fish and Reproduction success of lions. Pages 363-383 in T.H. Game, Juneau. 59pp. Clullon-Brock, ed. Reproductive success: studies of , AND J.L. HECHTEL. 1980. Structure, status, individual variation in contrasting breeding systems. Univ. reproductive biology, movement, distribution, and habitat Chicago Press, Chicago, Ill. 538pp. utilization of a grizzly bear population. Project W-17-l1. PAYNE, N.F. 1978. Hunting and management of the Job 4.14R, Alaska Dep. Fish and Game. Juneau. 65pp. Newfoundland black bear. Wildl. Soc. Bull. 6:206-211. __' AND __. 1984. Structure, status, reproductive PEARL, R. 1927. The growth of populations. Quart. Rev. biology, movement, distribution, and habitat utilization of BioI. 10:532-548. a grizzly bear population. Final Report. Project W-21-1, __, AND S.L. PARKER. 1922. Experimental studies on W-21-3, W-22-1, and W-22-2. Alaska Dep. Fish and the duration of life, IV: Data on the influence of density of Game, Juneau. 29pp. population on duration of life in Drosophila. Am. Nat. ROGERS, L.L. 1987a. Effects offood supply and kinship on 56:312-322. social behavior, movements, and population growth of 36 Int. Con! Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

black bears in northeastern Minnesota. Wild!. Monogr. Cleator, eds. Polynyas in the Canadian Arctic. Can. 97. 72pp. Wild!. Servo Occas. Pap. 45:7-27. . 1987b. Factors influencing dispersal in the black SMITH, R.B., AND L.1. VAN DAELE. 1991. Terror Lake bear. Pages 75-84 in B.D. Chepko-Sade and Z. T. Halpin, hydroelectric project, Kodiak Island, Alaska. Final report eds. Mammalian dispersal patterns. Univ. Chicago Press, on brown bear studies (1982-1986). Alaska Dep. Fish and Chicago, Ill. Game, Juneau. 188pp. ROFF, D.A. 1973. An examination of some statistical tests SMITH, T.G., AND I. STIRLING. 1978. Variation in the used in the analysis of mark-recapture data. Oecologia density of ringed seal (Phoca hispida) birth lairs in the 12:35-54. Amundsen Gulf, Northwest Territories. Can. J. Zool. ROYAMA, T. 1977. Population persistence and density 56:1066-1071. dependence. Ecol. Monogr.47: 1-35. SMITH, T .R. 1985. Ecology of black bears in a bottomland RUFF, R.L. 1982. Dynamics of black bear populations (low hardwood forest in Arkansas. Ph.D. Thesis, Univ. to no human exploitation). Proc. Western Black Bear Tennessee, Knoxville. 209pp. Workshop 2:87-103. STIRLING, I. 1974. Mid-summer observations on the RUSSELL, R.H., 1.W. NOLAN, N.G. WOODY, AND G.H. behavior of wild polar bears (Ursus maritimus). Can. J. ANDERSON. 1979. A study of the grizzly bear in Jasper Zool. 52:1191-1198. National Park, 1975 to 1978. Can. Wildl. Serv., , D. ANDRIASHEK, C. SPENCER, AND A. DEROCHER. Edmonton, AB. 136pp. 1988. Assessment of the polar bear population in the SCHAFFER, M.L. 1978. Determining minimum viable eastern Beaufort sea. Final report to the Northern Oil and population sizes; a case study of the grizzly bear (Ursus Gas Action Program by CWS, Edmonton, AB. 80pp. arctos). Ph.D. Thesis, Duke Univ. 192pp. __' __, ANDW. CALVERT. 1993. Habitat preferences __. 1983. Determining minimum viable population sizes of and distribution of polar bears in the western Canadian for the grizzly bear. Int. Conf. Bear Res. Manage. Arctic in late winter and spring. Polar Rec. 29: 13-24. 5:135-141. __, AND W.R. ARCHIBALD. 1977. Aspects of predation SCHOEN, 1., AND L. BEIER. 1990. Brown bear habitat of seals by polar bears. J. Fish. Res. Bd. Can. 34: 1126­ preferences and brown bear logging and mining 1129. relationships in southeast Alaska. Study 4.17. Alaska Dep. ____' AND D. DEMASTER. 1977a. Distribution and Fish and Game, Juneau. 90pp. abundance of seals in the eastern Beaufort Sea. Can. 1. SCHWARTZ, C.c., AND A.W. FRANZMANN. 1991. Zool. 34:976-988. Interrelationship of black bears to moose and forest , W. CALVERT, AND D. ANDRIASHEK. 1980. succession in the northern coniferous forest. Wildl. Population ecology studies of the polar bear in the area of Monogr. 113. 58pp. southeastern Baffin Island. Can. Wildl. Ser. Occas. Pap. , AND . 1992. Dispersal and survival of subadult No. 44. 33pp. black bears from the Kenai peninsula, Alaska. 1. Wildl. , AND A.E. DEROCHER. 1990. Factors affecting the Manage. 56:426-431. evolution and behavioral ecology of the modern bears. SCHWEINSBURG, R.E., L.1. LEE, AND P.B. LATOUR. 1981. Int. Conf. Bear Res. and Manage. 8:189-204. Distribution, movement and abundance of polar bears in ,c. JONKEL, P. SMITH, R. ROBERTSON, ANDD. CROSS. Lancaster Sound, Northwest Territories. Arctic 35:159­ 1977b. The ecology of the polar bear (Ursus maritimus) 169. along the western coast of Hudson Bay. Can. Wild\. Serv. SELLERS, R.A., AND L.D. AUMILLER. 1994. Brown bear Occas. Pap. 33. 64pp. population characteristics at McNeil River, Alaska. Int. , A.M. PEARSON, AND F.L. BUNNELL. 1976. Conf. Bear. Res. Manage. 9:In Press.. Population ecology studies of polar and grizzly bears in SERVHEEN, C. 1990. The status and conservation of the northern Canada. Trans. 41st North Am. Wild\. and Nat bears of the world. Eighth Int. Conf. Bear Res. and Resour. Conf. 41:421-431. Manage. Mono. Series No.2. STRATHEARN, S.M., G.B. KOLENOSKY, AND C.A. DEARY. SINCLAIR, A.R.E. 1977. The African buffalo. Univ. 1986. Movements and survival of orphan black bear cubs Chicago Press, Chicago, Ill. 355pp. in east-central Ontario. Int. Conf. Bear Res. and Manage. 1989. Population regulation in animals. Pages 7: (abstract only). 197-241 in 1.M. Cherrett, ed. Ecological Concepts. STRINGHAM, S.F. 1980. Possible impacts of hunting on the Blackwell Scientific Pub!. 385pp. grizzly/brown bear, a threatened species. Int. Conf. Bear SMITH, M.W., AND B.G. RIGBY. 1981. Distribution of Res. Manage. 4:337-349. polynyas in the Canadian Arctic. In I. Stirling and H. 1983. Roles of adult males in grizzly bear DENSITY EFFECTS ON BEARS • Taylor et al. 37

population biology. Int. Conf. Bear Res. Manage. TIETJE, W.D., B.O. PELCHAT, AND R.L. RUFF. 1986. 5: 140-151. Cannibalism of denned black bears. 1. Mammal. 67:762­ 1985. Responses by grizzly bear population 766. dynamics to certain environmental and biosocial factors. __' AND R.L. RUFF. 1983. Responses of black bears to Ph.D. Thesis, Univ. Tenn., Knoxville. 456pp. oil development in Alberta. Wildl. Soc. Bull. 11 :99-112. __. 1986. Effects of climate, dump closure, and other TROYER, W.A., AND R.I. HENSEL. 1962. Cannibalism in factors on Yellowstone grizzly bear litter size. Int. Conf. brown bear. Anim. Behav. 10:231. Bear Res. Manage. 6:33-39. VERHULST, P.F. 1838. Notice sur la loi que la population ___. 1990a. Grizzly bear reproductive rate relative to pursuit dans son accroissement. Corresp. Math. Phys. body size. Int. Conf. Bear Res. Manage. 8:433-443. 10:113-121. __. 1990b. Black bear reproductive rate relative to body VICKERY, W.L., AND T.D. NUDDS. 1984. Detection of weight in hunted populations. Int. Conf. Bear Res. and density dependent effects in annual duck censuses. Manage. 8:425-432. Ecology 65:96-104. STRONG, D.R. 1986. Density-vague population change. WATTS, P.D., AND S.E. HANSEN. 1987. Cyclic starvation Trends Eco!. Evol. 1:39-42. as a reproductive strategy in the polar bear. Symp. Zoo!. __. 1987. The density dependence debate: reply from Soc. London. 57:305-318. D.R. Strong. Trends Eco!. Evol. 2:76-77. WOODING, I.B. ,AND T.S. HARDISKY. 1994. Home range TAYLOR, M.K., D. DEMASTER, F.L. BUNNELL, AND R. and habitat use of black bears in north central Florida. SCHWEINSBURG. 1987a. Modelling the sustainable harvest Int. Conf. Bear Res. and Manage. 9:In Press. of female polar bears. 1. Wild!. Manage. 51:811-820. YODZIS, P., AND G.B. KOLENOSKY. 1986. A population __' __, __' __' AND SMITH, 1. 1987b. dynamics model of black bears in eastcentral Ontario. 1. ANURUS: a population analysis system for polar bears. Wild!. Manage. 50:602-612. Int. Conf. Bear Res. and Manage. 7:117-125. YOUNG, B.F., AND R.L. RUFF. 1982. Population dynamics __, T. LARSEN, AND R. SCHWEINSBURG. 1985. and movements of black bears in east central Alberta. J. Observation of intraspecific aggression and cannibalism in Wildl. Manage. 46:845-860. polar bears (Ursus maritimus). Arctic. 38:303-309. YOUNG, E.L. 1990. Game Management Unit 4. Pages 7-20 , AND J.L. LEE. Submitted. The distribution and in S.O. Morgan, ed. Annual Report of Survey-Inventory abundance of Canadian polar bear populations. Arctic. Activities, January-December 1988, Fed. Aid in Wildl. Submitted. Restor. Project W-23-2, Study 4.0, Vol. 20 part 5. Alaska Dep. of Fish and Game. 189pp. 38 Int. Cont Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

Appendix I. Density-Dependent Population Regulation and Harvest Policies for Bears

The question of population regulation of bears has a population growth rate is central to population long and distinguished history. Initial controversy on dynamics, competition, predation, speciation, the regulation of natural populations was one of the community structure theory, and conservation of primary forces that resulted in the development of harvested populations. Whether harvest mortality is Ecology as a formal discipline. The first treatment of additive or compensatory for bear populations is only a this issue was intended for human populations (Malthus rephrasing of the initial debate that launched the 1798). Malthus (1798) identified war, famine, and discipline of Ecology. At face value, the answer would disease as "positive checks" and homosexuality, birth seem to be trivial. Bear populations like any other control, delayed marriage, and abstinence as populations are ultimately regulated by density­ "preventative checks" that limit population growth. The dependent factors. But several important details remain relevance of preventative checks to bear populations is uncertain. How does density act to regulate a given obviously limited, however, nutrition and intraspecific population of bears: is the feedback on survival or aggression are the main factors thought to regulate bear reproductive rates or both? Is there a critical densities. population level required before density effects begin? A mathematical equation describing population Do density effects reduce survival and recruitment growth to some equilibrium density (carrying capacity) parameters in a linear or nonlinear manner? Are all sex was developed by Verhulst (1838) and later termed the and age strata affected, or only certain ones? The "Logistic Equation" by Pearl (1927). Charles Darwin mechanism of density effects will determine the (Chapter 3, 1859) noted that populations produced relationship between density of animals and population­ considerably more individuals than their habitat would growth rate, and the response of sustainable harvest support. This surplus of individuals (types) was the rates to population reductions. To envision these raw material for natural selection. Modern ecological dynamics it is useful to revisit the equations for thinking about the role of density in changes in population growth in discrete time developed by Pearl population numbers was begun by entomologists (1927): (Howard and Fiske 1911). The order and simplicity of Nt+1=Nt+N/(B-D) the logistic equation was supported by several (1) N + I =Ntx A laboratory studies (Pearl and Parker 1922, Gause 1934, t Bodenhiemer 1938). Simultaneously, outbreaks of pest insects were found to be more related to physical When there is a linear decline in birth rate (B) or a conditions (e.g., weather) than to previous or existing linear increase in death rate (D) or both with increasing densities (Andrewartha and Birch 1954). Conversely, numbers (Fig. 7), a specific type of density-dependent Lack (1967) documented the close correspondence of population regulation occurs. many bird populations to the predictions of logistic B=B -SBxN limitation of population numbers. Probably the most max (2) D=D +SD xN vocal proponent of density effects as a regulating factor min in population density was Nicholson (1954a,b).

The debate over whether population numbers were Where: Bmax = maximum birth rate,

limited by physical (density-independent) or density­ Dmill = minimum death rate, dependent effects was resolved at the Cold Spring SB = slope of B versus N line, and Harbour Symposium on Quantitative Biology held in SD = slope of D versus N line. 1957. Although many populations may exist for long For convenience allow ~ = B-D and A = 1+~, then periods at densities where density-dependent factors are equation 3 may be rewritten in terms of equations 2. If insignificant relative to density-independent factors, density affects births only: logically, density-independent factors can not regulate I1=Bmax-SBxN-D populations indefinitely. Factors affecting demography (3) = (B -D)-SB xN may be functionally separated into limiting (affects max population growth equally at all densities) and regulatory (effects increase with increased densities). Only regulatory effects are density dependent. The concept of density induced limits to DENSITY EFFECTS ON BEARS • Taylor et al. 39

BMAJ(

POPULATION NUMBER (N) • POPULATION NUMBER (N)..

Fig. 7. The simplest form of density effects occur when increasing population numbers cause a linear decrease in per capita birth rate or a linear decrease in per capita death rate.

If density affects deaths only: The effect on various forms of density effects on the sustained yield of a harvested population may be Ll=B-[D -(SD>«A-l)>

Ll=(B -SB>

Substitution for J... in equation 1 yields the familiar (9) logistic equation: N >«l-~)] N /.] =N{ ><11 +Ll max K (7) 40 Int. Conf Bear Res. and Manage. Monogr. Ser. No.3, 1994. 43pp.

300 1.2 6 MAX = BMAl( - ° 01' = B - 0MIN 'rl. • MAX_ 01' =BMAX-OMIN :; g III 200 0.8 ~ 6 Iii a: 6=6 - MAxl. N &: (+) MAX r[ K ]

(-) L--'-_'------'------'_--'------'-_--'-----'-_--'--'" 0.0 200 400 600 800 "K." POPULATION NUMBER (N) K = Carrying Capacity = Density lit which 6 = 0 or > >"=1.0 Fig. 9. When the birth rate - death rate (11) is a decreasing linear function of population number, the maximum Fig. 8. When increasing population numbers cause a linear sustainable harvest rate is parabolic with a maximum at % decrease in birth rate (B) or a linear increase in death rate (D) carrying capacity (K). See equation 9 in Appendix I for or both, then 11 (B-D) is also a decreasing linear function of details. population number.

The relationship of maximum sustainable harvest to monality are discrete categories, however when the population density is parabolic, with the maximum effects of density are felt across the entire range of occurring at 1/2 K (Fig. 9). This "logistic" example, population densities, as in the logistic model, harvest which is based on the assumption of linear relationships monality is a continuum of additive and compensatory between binh and death with population numbers, has monality. At population densities close to K, harvest caused some managers to conclude that all populations should be reduced to 112 K for maximum sustainable yield. However, there is compelling evidence that long-lived species with low reproduction potential are not affected by density in a linear manner (Eberhardt .t:. MAX 1------,. and Siniff 1977, Fowler 1981a,b, 1990). A simple form of nonlinear density effects is a step g function, where harvest monality is strictly additive at w ~ low densities, and compensatory monality occurs at a: higher densities (Fig. 10). Compensatory harvest ::r monality is "compensated for" by reduced natural !:iw Q mortality or increased recruitment. Additive harvest I mortality is added to natural monality for total W !:i a: mortality. Actual harvest mortality probably never falls Additive Compensatory totally within either idealized category. To be strictly ::r Harvest Harvest ~ additive, hunters would take only those animals that iii Mortality Mortality would have survived without the harvest. Also 0'------'------''''''-­ P9pulation densities would have to be sufficiently low o K POPULATION NUMBER (N) that no density effects were occurring. Compensatory > mortality can result in a sustainable harvest that is largest at some intermediate density (e.g., from V2K in Fig. 10. This hypothetical step function provides an example the logistic model). The step function (Fig. 10) of density effects where additive and compensatory mortality suggests that additive and compensatory harvest would exist as distinct categories. DENSITY EFFECTS ON BEARS • Taylor et at. 41 mortality is mostly compensatory and at population survival or recruitment rate. The rate term could also densities close to 0 harvest mortality is mostly additive. be per-capita rates of increase (.6.) in which case CC The step function in Figure 10 suggests that any would be the carrying capacity (K). population in which harvest mortality is additive, is A nonlinear form of density effects gives a being harvested beneath maximum sustainable yield sustainable harvest function that is skewed towards (i.e., if the manager allowed the population to increase carrying capacity (Fig. 12). Populations which have to the point that it began to experience density effects, nonlinear density functions (e.g., bears, elephants, and a larger harvest would be possible). marine mammals) will sustain the highest harvest rates A more realistic density function than the step at densities close to carrying capacity. The maximum function is a continuous nonlinear curve (Eberhardt and sustained yield will depend on the degree of Siniff 1977, Fowler 1981a,b, 1990). An equation that nonlinearity in the density effects. Documenting the is useful for describing nonlinear density effects is: mechanisms of density effects will require long-term studies of all vital rates over the range of population [( CC)_I] densities that density effects occur. Once populations N R=J\na x _ (10) are reduced below that critical range, fluctuations in the x [( CC)-l+KS] per capita rate of increase will be mainly density N independent and harvest mortality will be mainly additive. This equation (Fig. 11) is an inverted form of the Although the shape of the density effects curve is not Mechalis-Menton equation used to describe enzyme known for any species or population of bears, the kinetics. The KS term controls the degree of possibilities for maximum sustained yield may be nonlinearity in the density relationship. At KS = 1.0, bounded by considering the possible ranges of density the logistic curve is recovered. The value CC is the effects. Assuming for convenience, a carrying capacity population density where the survival or recruitment of 1,000, the range of maximum sustained yields rate equals zero. The Rmat value is the maximum between logistic density effects (KS = 1) and highly

1.0 - ~ '-'":::::::::.------_ --- -- a: ­ - 0.8 - "- - "­ - KS=O.10 ...... , ~ i "" "LOGISTIC" " , a: KS = 1.00 ", 0.6 \ u.. \ \ o \ ' \ z \ 0.4 \ o \ 0 \ I-- \ o \ , \ ' « 0.2 \ : a: \ : CARRYING \: u.. CAPACllY \: (CC) \: ~ OL.J....-.....I----'----'----'--.....I----'--...... -....L----'----.,;-.... o 200 400 600 800 1000 ABUNDANCE (N)

Fig. 11. A useful equation for describing both linear (logistic) and nonlinear density effects was developed by inverting the threshold corrected Mechalis-Menton equation. This formulation has the advantage of flexibility and easily understood parameters. The KS term controls the degree of nonlinearity. At KS = 1.0, logistic density effects are recovered.