There is considerable controversy in the literature about the presence of density dependence in dispersal. In this study, we exploit a data series from a long-term study (>18 years) on radio-marked brown bears (Ursus arctos L.) in two study areas in Scandinavia to investigate how individual-based densities influence the probability of natal dispersal and natal dispersal distances. Cumulatively, 32% and 46% of the females and 81% and 92% of the males dispersed before reaching 5 years of age in the northern and southern study area, respectively. Density had a negative effect on both the probability of dispersal and dispersal distances for the dispersing animals, when controlling for study area, sex and age, making this the first study to show that natal dispersal probability and distances are inversely density dependent in a large carnivore. We suggest that female-female competition for space caused females in higher density areas to settle closer to their natal area. For males, however, merging of demes, resulting in decreased relatedness and increased heterozygosity in an expanding population, might be the reason for shorter dispersal distances in males living at higher densities. This has been hypothesised for small mammals. The high proportion of dispersing female brown bears in Scandinavian compared with North American studies might be due to lower densities in Scandinavia and recent population expansion, with unoccupied areas available at the edges of the population. The longer dispersal distances in female Scandinavian brown bears suggest less social constraints on movements than for North American females. The longer dispersal distances by Scandinavian males may be due to increased searching for potential mates in peripheral areas with lower densities of females. These results, in addition to results of other brown bear studies, suggest that brown bears might be more territorial than previously thought, and that density is regulated by social interactions.

Habitat selection is a fundamental behaviour that links individuals to the resources required for survival and reproduction. Although natural selection acts on an individual's phenotype, research on habitat selection often pools inter-individual patterns to provide inferences on the population scale. Here, we expanded a traditional approach of quantifying habitat selection at the individual level to explore the potential for consistent individual differences of habitat selection. We used random coefficients in resource selection functions (RSFs) and repeatability estimates to test for variability in habitat selection. We applied our method to a detailed dataset of GPS relocations of brown bears (Ursus arctos) taken over a period of 6 years, and assessed whether they displayed repeatable individual differences in habitat selection toward two habitat types: bogs and recent timber-harvest cut blocks. In our analyses, we controlled for the availability of habitat, i.e. the functional response in habitat selection. Repeatability estimates of habitat selection toward bogs and cut blocks were 0.304 and 0.420, respectively. Therefore, 30.4 and 42.0 % of the population-scale habitat selection variability for bogs and cut blocks, respectively, was due to differences among individuals, suggesting that consistent individual variation in habitat selection exists in brown bears. Using simulations, we posit that repeatability values of habitat selection are not related to the value and significance of β estimates in RSFs. Although individual differences in habitat selection could be the results of non-exclusive factors, our results illustrate the evolutionary potential of habitat selection.

No abstract