A predator's species, sex and body size can influence the types of prey that it consumes, but why? Do such dietary divergences result from differences in foraging habitats, or reflect differential ability to locate, capture or ingest different types of prey? That question is difficult to answer if foraging occurs in places that preclude direct observation. In New Caledonia, amphibious sea kraits (Laticauda laticaudata and L. saintgironsi) mostly eat eels—but the species consumed differ between snake species and vary with snake body size and sex. Because the snakes capture eels within crevices on the sea floor, it is not possible to observe snake foraging on any quantitative basis. We used stable isotopes to investigate habitatdivergence and ontogenetic shifts in feeding habits of sympatric species of sea kraits. Similarities in d15 N (*10.5%) values suggest that the two snake species occupy similar trophic levels in the coral-reef foodweb. However, d13C values differed among the eight eel species consumed by snakes, as well as between the two snake species, and were linked to habitat types. Specifically, d13C differed between soft- vs. hard-substrate eel species, and consistently differed between the soft-bottom forager L. laticaudata (* -14.7%) and the hard-bottom forager L. saintgironsi (* -12.5%). Differences in isotopic signatures within and between the two sea krait species and their prey were consistent with the hypothesis of habitatbased dietary divergence. Isotopic composition varied with body size within each of the snake species and varied with body size within some eel species, reflecting ontogenetic shifts in feeding habits of both the sea kraits and their prey. Our results support the findings of previous studies based on snake stomach contents, indicating that further studies could usefully expand these isotopic analyses to a broader range of trophic levels, fish species and spatial scales.