Differences in diet composition among conspecifics (dietary specialization) have been documented across a broad range of taxonomic groups and habitats, and such variation at the individual level is increasingly recognized as an important component of diversity in trophic interactions. Accurate identification of individual dietary specialization, however, requires longitudinal dietary records that are labor-intensive and cost-prohibitive to obtain for many species. Here we explore the use of stable isotopes (delta13C and delta15N) as a promising technique for detecting and quantifying patterns of individual dietary specialization. Southern sea otters (Enhydra lutris nereis) offer a unique opportunity for testing this approach because (1) they consume a wide variety of prey that span multiple trophic levels, habitats, and ecologically defined functional groups; and (2) individual diet specialization can be validated with existing observational data. We analyzed the isotopic composition of sea otter vibrissae (n = 31) in order to characterize inter- and intra-individual variation in sea otter diets at Monterey Bay, California, USA. At the population level, sea otters showed substantial variation in both delta13C and delta15N values, occupying nearly all of the "isotopic space" created by the diversity of isotopic signatures of potential prey taxa. Most of the variation in sea otter vibrissae was accounted for by differences between individuals, with much less contributed by within-individual variation. A majority of sea otters (approximately 80%) showed relatively little temporal variability in isotopic composition, suggesting that the proportional composition of most individuals' diets is relatively constant over time; a few individuals (approximately 20%) exhibited a high degree of intra-vibrissa isotopic variability, suggesting seasonal shifts in diet composition. These results and our interpretation of them were supported by long-term observational data on the diets of radio-tagged sea otters from the same population (n = 23). Our results demonstrate that stable isotopes can provide an efficient tool for measuring individual- and population-level dietary breadth and may be useful for studying populations where longitudinal data on individuals would otherwise be impossible to acquire. This will be critical for examining the causes and consequences of dietary variation within and among consumer populations, thereby improving our understanding of these important ecological and evolutionary processes at the community level.