In our study, male cougars had higher inbreeding coefficients than females across all sites, and although differences were not statistically significant with the exception of the Blue Mountains cougars, we observed an east-to-west gradient, which was particularly pronounced for males in the coastal regions, especially on the Olympic Peninsula. This may be an indication that gene flow of male cougars is more limited across these human-altered landscapes, which corroborates the findings of Zeller et al. (2023) reporting that male cougars had a higher resistance to movement across developed, built-up areas when compared to their female counterparts.

Conversely in areas with higher human-caused mortality such as the northern Cascades and northern Rocky Mountains, cougar genetics were archetypal of having more immigration than emigration. Delibes et al. (2001) and Robinson et al. (2008) described this kind of site as an ‘attractive sink’ that has high-quality habitat and abundant resources, but also increased levels of human-caused mortality. To fill unoccupied territories after the numbers of resident cougars are reduced, dispersing subadults emigrate from adjacent areas into these vacant areas (e.g., Logan & Sweanor 2001; Robinson et al. 2008).

Comparative diversity analysis of the geographic regions revealed that the genetic diversity was highest for cougars sampled in the Northern Rocky Mountains region (HE = 0.58) and lowest for cougars on the Olympic Peninsula (HE = 0.47) (Table 1), but differences between sites were not statistically significant (Kruskal Wallis Test, H = 2.34, df = 5, P = 0.800).

Extraction and PCR negatives were added to all reactions to control for contamination. PCR products were visualized using Gene-Scan 500 LIZ sizing standard (Applied Biosystems™, Carlsbad, CA) and an ABI 3730 DNA analyzer (Applied Biosystems™, Carlsbad, CA). Alleles were scored using GENEMAPPER, version 3.7 (Applied Biosystems™, Carlsbad, CA). DNA extraction and genotyping were conducted at the WDFW Molecular Genetics Laboratory in Olympia, WA, US. We used R package AlleleMatch, version 2.5.1 (Galpern et al. 2012) to identify individual multilocus genotypes and recaptures. We also confirmed unique identities of cougar genotypes by calculating probabilities of identity between siblings (P(ID)sibs) using Gimlet, version 1.3.3 (Valière 2002), as recommended by Mills et al. (2000) and Waits et al. (2001). Deviations from Hardy–Weinberg equilibrium (HWE) and linkage disequilibrium (LD) were assessed using exact tests and Markov chain Monte Carlo (MCMC) estimation (iterations per batch = 5,000; dememorization = 10,000; batches = 5) in the R package genepop, version 1.1.4 (Rousset 2008). To test the random linkage or association of loci, we also calculated the standardized index of association (rD, Brown et al. 1980) using a permutation approach (n = 999). Significant levels of multiple comparisons were adjusted by applying a sequential Bonferroni correction (Rice 1989). In addition, we also screened all microsatellite loci for the occurrence of null alleles using MICROCHECKER, version 2.3.3 (Van Oosterhout et al. 2004).