A simulation method was developed for identifying populations with levels of human‐caused mortality that could lead to depletion, taking into account the uncertainty of available information. A mortality limit (termed the Potential Biological Removal, PBR, under the U. S. Marine Mammal Protection Act) was calculated as the product of a minimum population estimate (NMIN), one‐half of the maximum net productivity rate (RMAX), and a recovery factor (FR). Mortality limits were evaluated based on whether at least 95% of the simulated populations met two criteria: (1) that populations starting at the maximum net productivity level (MNPL) stayed there or above after 20 yr, and (2) that populations starting at 30% of carrying‐capacity (K) recovered to at least MNPL after 100 yr. Simulations of populations that experienced mortality equal to the PBR indicated that using approximately the 20th percentile (the lower 60% log‐normal confidence limit) of the abundance estimate for NMIN met the criteria for both cetaceans (assuming RMAX= 0.04) and pinnipeds (assuming RMAX= 0.12). Additional simulations that included plausible levels of bias in the available information indicated that using a value of 0.5 for FR would meet both criteria during these “bias trials.” It is concluded that any marine mammal population with an estimate of human‐caused mortality that is greater than its PBR has a level of mortality that could lead to the depletion of the population. The simulation methods were also used to show how mortality limits could be calculated to meet conservation goals other than the U. S. goal of maintaining populations above MNPL.

Killer whales (Orcinus orca) currently comprise a single, cosmopolitan species with a diverse diet. However, studies over the last 30 yr have revealed populations of sympatric “ecotypes” with discrete prey preferences, morphology, and behaviors. Although these ecotypes avoid social interactions and are not known to interbreed, genetic studies to date have found extremely low levels of diversity in the mitochondrial control region, and few clear phylogeographic patterns worldwide. This low level of diversity is likely due to low mitochondrial mutation rates that are common to cetaceans. Using killer whales as a case study, we have developed a method to readily sequence, assemble, and analyze complete mitochondrial genomes from large numbers of samples to more accurately assess phylogeography and estimate divergence times. This represents an important tool for wildlife management, not only for killer whales but for many marine taxa. We used high-throughput sequencing to survey whole mitochondrial genome variation of 139 samples from the North Pacific, North Atlantic, and southern oceans. Phylogenetic analysis indicated that each of the known ecotypes represents a strongly supported clade with divergence times ranging from ∼150,000 to 700,000 yr ago. We recommend that three named ecotypes be elevated to full species, and that the remaining types be recognized as subspecies pending additional data. Establishing appropriate taxonomic designations will greatly aid in understanding the ecological impacts and conservation needs of these important marine predators. We predict that phylogeographic mitogenomics will become an important tool for improved statistical phylogeography and more precise estimates of divergence times.

We quantified the relative influence of maternal fidelity to feeding grounds and natal fidelity to breeding grounds on the population structure of humpback whales Megaptera novaeangliae based on an ocean-wide survey of mitochondrial (mt) DNA diversity in the North Pacific. For 2193 biopsy samples collected from whales in 10 feeding regions and 8 breeding regions during the winter and summer of 2004 to 2006, we first used microsatellite genotyping (average, 9.5 loci) to identify replicate samples. From sequences of the mtDNA control region (500 bp) we identified 28 unique haplotypes from 30 variable sites. Haplotype frequencies differed markedly among feeding regions (overall F ST = 0.121, Φ ST = 0.178, p < 0.0001), supporting previous evidence of strong maternal fidelity. Haplotype frequencies also differed markedly among breeding regions (overall F ST = 0.093, Φ ST = 0.106, p < 0.0001), providing evidence of strong natal fidelity. Although sex-biased dispersal was not evident, differentiation of microsatellite allele frequencies was weak compared to differentiation of mtDNA haplotypes, suggesting male-biased gene flow. Feeding and breeding regions showed significant differences in haplotype frequencies, even for regions known to be strongly connected by patterns of individual migration. Thus, the influence of migratory fidelity seems to operate somewhat independently on feeding and breeding grounds over an evolutionary time scale. This results in a complex population structure and the potential to define multiple units to conserve in either seasonal habitat.

Analysing population genomic data from killer whale ecotypes, which we estimate have globally radiated within less than 250,000 years, we show that genetic structuring including the segregation of potentially functional alleles is associated with socially inherited ecological niche. Reconstruction of ancestral demographic history revealed bottlenecks during founder events, likely promoting ecological divergence and genetic drift resulting in a wide range of genome-wide differentiation between pairs of allopatric and sympatric ecotypes. Functional enrichment analyses provided evidence for regional genomic divergence associated with habitat, dietary preferences and post-zygotic reproductive isolation. Our findings are consistent with expansion of small founder groups into novel niches by an initial plastic behavioural response, perpetuated by social learning imposing an altered natural selection regime. The study constitutes an important step towards an understanding of the complex interaction between demographic history, culture, ecological adaptation and evolution at the genomic level.