Anatomy of an invasive metapopulation: The California nonnative red fox

Background/Question/Methods

Invasive species can have detrimental effects on native communities, including decimation of threatened or endangered prey, competition with native species, or maladaptive hybridization with closely related taxa. Understanding the processes of invasions and factors maintaining them is therefore important. We studied this problem with red foxes (Vulpes vulpes). The rise and fall of the 20^th century red fox fur-farming industry entailed releases of captive-reared foxes, some of which founded invasive populations. Over the past 5 decades, such populations in California expanded from one or more foci to their current, relatively continuous distribution over most low-elevation portions of California; the exception was the Sacramento Valley, where native red foxes already occurred (V. v. patwin). Nonnative foxes impact numerous endangered ground-nesting bird species and threaten the genetic integrity of the native Sacramento Valley red fox through hybridization. We used a genetic approach involving 14 microsatellites and mitochondrial DNA (mtDNA) sequences of 294 nonnative red foxes to investigate initial sources, routes of spread, contemporary connectivity, and potential source-sink dynamics. We employed admixture analysis in Program Structure, genetic distance measures, estimates of genetic diversity, and bottleneck tests to assess source populations, connectivity, and potential demographic source and sink populations.

Results/Conclusions

Admixture analysis parsed nonnative red fox populations into two distinct clusters: the San Francisco Bay (SFB) salt marsh populations and all other populations (nonSFB). Within these genetic clusters, we detected additional substructure. The SFB cluster was subdivided into 2 adjacent subpopulations, each corresponding to a different dominant mtDNA haplotype. Genetic diversity in SFB was highest in the easternmost subpopulation and lowest in the westernmost subpopulation, suggesting east-to-west colonization. The nonSFB cluster was subdivided into Central-Coast, Southern California, and San Joaquin Valley subpopulations, each of which also was characterized by a distinct dominant mtDNA haplotype. Genetic diversity was highest in these three regions, with intervening areas showing lower diversity. Genotypes from intervening locations also reflected admixture from adjacent sources and genetic distance (Da) measures exhibited an isolation-by-distance relationship (r ²=0.43, P < 0.001). Together, these patterns were most consistent with expansions from three distinct sources of introduction. Bottleneck analyses showed evidence of a population expansion in the SFB area but no evidence of expansion or population decline (e.g., indicative of a demographic sink) in any other population. This is consistent with a network of self-sustaining populations linked by, but not dependent on, migration.