The extent to which community processes can be generalized from local field studies to larger spatial scales remains a contentious issue. The search for broad generality can be hampered when species interactions vary geographically, a common phenomenon attributed to a wide range of underlying ecological factors. Less attention has been directed toward understanding the additional role that evolutionary processes may play in modifying the way that pairs of species interact over large spatial scales. Here we examine whether marked geographic variation in the interaction between a predatory snail (Nucella canaliculata) and an intertidal mussel (Mytilus californianus) arises from phenotypic plasticity or fixed genetic differences among snail populations. Over a three-year period, we reared snails from eight populations in California and Oregon, USA, through two laboratory generations and tested whether family lines differed in their ability to drill M. californianus. Remarkably, F2 generation snails from Oregon sources were generally unable to drill mid-sized Mytilus californianus (5-7 cm long), whereas snails from California readily drilled this prey. Because snails were raised through two generations on a common diet (Mytilus trossulus), these differences among populations likely have a genetic basis. Snails from California and Oregon readily interbred and produced viable offspring in the laboratory, suggesting that populations belong to a single biological species. Field surveys of mussel beds revealed striking geographic variation in predation that closely matched the observed differences in the drilling capacity of N. canaliculata. Drilled M. californianus were common at all sites in California and included many large mussels (> 10 cm long). In contrast, drilled M. californianus were rare on the central Oregon coast and consisted mostly of small mussels (< 3 cm). We hypothesize that persistent variation in prey recruitment along the coast has selected for interpopulation differences in the drilling capacity of this direct-developing snail, with potential consequences for the size structure of mussel beds. Combined with growing evidence of restricted gene flow and low connectivity among many marine populations, this study highlights the importance of considering the contribution of evolutionary processes to geographic variation in species interactions.