Overall, the authors did an outstanding job applying population genetics techniques appropriately. In particular, the authors made very good use of the program STRUCTURE. This program was used to detect population substructure and if detected, the largest genetic cluster was selected for inclusion in ensuing analyses. This important step prevents false detection of bottlenecks, which can be a common mistake. It is also appreciated that the authors chose to examine allelic richness rather than allelic frequency. We were left with a few lingering questions about the methods, however. First, we are curious if the authors acquired raw electropherograms or pre-interpreted genotypes for the published datasets used and if there were any measures taken to control for observer bias in interpreting microsatellite genotypes, such as preparing and running samples from other labs and assessing whether similar conclusions were reached. We were also curious about the justification for grouping IUCN categories into "concern" ('near threatened,' 'vulnerable,' and 'endangered') and "least concern" (l'east concern'), especially since the IUCN Red List Categories and Criteria groups 'near threatened' with 'least concern' and explicitly distinguishes 'vulnerable,' 'endangered,' and 'critically endangered' as the "threatened" categories. We wonder how the results presented in Figure 6 would be affected by moving the species designated as 'near threatened' out of the "concern" category.
Lastly, we would have appreciated a more extensive discussion. We felt it was a large assumption that marine/ice-breeding habitats are less disturbed. In the 1800's, this may have been the case although indigenous anthropogenic hunting was probably already occurring (albeit on a smaller scale, ref). However this assumption needed some justification in the introduction. Even more importantly, the implications of the change in anthropogenic disturbance patterns and continuing disturbance  in these habitats into the future need to be addressed in the discussion.