The ratio of N-15/N-14 (delta N-15) from consumer and prey tissue is commonly used in ecological studies to determine trophic level, food web structure, and mean trophic level in aquatic ecosystems. There is a predictable positive relationship between the delta N-15 values in tissue and trophic level, caused by the bioaccumulation of N-15 in tissues of consumers with each step up the food chain. Reconstructing trophic structure or food chain length over time may provide resource managers with insights about ecosystem biodiversity and resilience. Yet, in many marine systems the absence of baseline information before anthropogenic disturbances makes comparative studies addressing ecosystem responses extremely difficult. Here we attempt to retrospectively reconstruct trophic position in four species of fish from the upper Gulf of California, Mexico before perturbations such as overfishing or the damming of the Colorado River. We first validated if otolith delta N-15 approximates the delta N-15 observed in fish tissue. We then used the delta N-15 encapsulated in ancient fish otoliths that are between 1,000 and 5,500 years old to define the food web structure. Our results suggested that delta N-15 in otoliths has slightly more positive delta N-15 than soft tissue. The delta N-15 values from ancient otoliths appropriately defined the fishes' relative trophic position. We found significant differences in delta N-15 between functional groups, apex predator versus intermediate predators. Juveniles and adult fishes displayed trophic separation between functional groups. Our findings advocate the application of delta N-15 analysis of prehistoric otoliths for establishing pre-disturbance ecological benchmarks.