Abstract. Lodgepole pine dwarf mistletoe (Arceuthobium americanum), an aggressive parasite of lodgepole pine (Pinus contorta subsp. latifolia), is limited, in the Central Rocky Mountains, to montane host "islands." We sought to quantify the effect of the insularity and life history of A. americanum on its population genetic structure by considering three questions: (1) Do apparent limits to A. americanum pollen and seed dispersal affect gene flow at small spatial scales? (2) Are patterns of genetic diversity on montane "islands" and "mainlands" consistent with the predictions of island biogeographic theory? (3) What is the origin and genetic relationship of "island" and "mainland" populations of A. americanum within the Central Rocky Mountain study area? We used seven A. americanum microsatellite markers developed specifically for the project to genotype 241 individuals from six populations (four "island" and two "mainland" locations) at the western edge of the Central Rocky Mountains, Idaho, USA. All loci were unlinked and polymorphic with a range of 4-24 alleles per locus. To consider gene flow, we performed spatial analyses and compared F-statistics from our data to those from a wide range of species and functional groups. We tested the predictions of biogeographic theory by quantifying associations of genetic diversity with "island" area and isolation from "mainland" populations. To consider the genetic relationships of "island" and "mainland" regions, we applied agglomerative and Bayesian hierarchical classifications to datasets, and objectively identified the optimal numbers of clusters in these analyses. We found that F IS values in A. americanum were similar to those reported for dispersal-limited species with similar life-history characteristics, and F ST values were similar to those reported for plants with animal-facilitated seed dispersal. As predicted by biogeographic theory, island allelic richness was positively correlated with geographic size of the population and negatively correlated with distance from mainland populations. Average linkage (UPGMA) classifications and Bayesian cluster analyses both identified two groupings of island populations radiating from the mainland populations. Our results are likely to be of interest to foresters managing dwarf mistletoe infections, particularly in the Central Rocky Mountains, and to ecologists and geneticists working with insular population distributions.