We investigated the use of geometric morphological shape differences to differentiate laboratory-reared larval (4–22 days posthatch, < 10 mm SL) Morone chrysops, Morone saxatilis, and Morone chrysops ♀ × M. saxatilis ♂ hybrids. We also examined impacts of allometry on descriptions of individual shape. For validation of a shape-based taxonomic discrimination model, we used cellulose acetate electrophoresis to establish a “known species” test group of field-collected larvae based on banding patterns for the enzyme system esterase. Geometric shape was described with Cartesian coordinates of 16 anatomical landmarks located along the midsaggital outline of laboratory-reared (n = 373) and field-collected (n = 29) larvae. Coordinate data were reoriented and rescaled to uniform centroid size and analyzed as landmark displacements from a reference form. Discriminant function analysis resulted in 100% separation of M. chrysops and M. saxatilis larvae based on geometric shape. Discrimination success dropped to 87% when more variable data from hybrids were included in the discriminant function. Seventy percent and 87%, respectively, of field collected larvae (n = 29) and a randomly selected laboratory subgroup (n = 40) were correctly classified to taxon. Results suggest that taxonomic separation of early-stage larvae based on shape data is not affected by allometry but is sensitive to larval nutritional condition and handling. Landmark-based morphometrics may provide an improved, simple method of discrimination among species in sympatric populations of closely related larval fishes.