Here we examine genetic characteristics of brown bears of Kodiak and Afognak islands, using 14 variable nuclear microsatellite loci and nucleotide sequence information including the hypervariable domain I of the mtDNA control region (Wakely 1993). Because these markers, or a subset of them, have been used to characterize brown bears of the Kenai Peninsula (Jackson et al. 2005), Katmai National Park, Seward Peninsula, and nine other populations in Alaska (Talbot, unpublished data), we compared levels of genetic diversity and relationships among populations when possible. In addition, we obtained preliminary comparative information from class II DQA and DQB genes of the brown bear MHC, to examine levels of variation at this important immunology-mediating supergene. These data were used to answer the following questions: 1) are earlier findings of extremely low levels of variability at nuclear (biparentallyinherited) microsatellite loci from a small geographic area (Paetkau et al. 1998b) representative of Kodiak Archipelago populations as a whole? 2) Is the level and type of variation at the maternally-inherited mtDNA lower, or similar to, levels found in other populations in Alaska? 3) Is there concordance between low levels of genetic variation observed at neutral markers with levels of variation observed at functional genes? 4) Is there population substructuring within Kodiak and Afognak islands? 5) What is the connectivity between populations on Afognak Island and Kodiak Island? 6) What are the phylogeographic relationships between bears of the Kodiak Archipelago with brown bears on mainland Alaskan and other western Beringian populations? We also test whether these markers will provide an appropriate baseline for designing genetic tagging studies for use in future research and management activities, such as mark-recapture efforts, on the Refuge.