Mobility is essential to the fitness of many animals, and the costs of locomotion can dominate daily energy budgets. Locomotor costs are determined by the physiological demands of sustaining mechanical performance, yet performance is poorly understood for most animals in the field, particularly aquatic organisms. We have used 3‐D underwater videography to quantify the swimming trajectories and propulsive modes of bluegills sunfish (Lepomis macrochirus, Rafinesque) in the field with high spatial (1–3 mm per pixel) and temporal (60 Hz frame rate) resolution. Although field swimming trajectories were variable and nonlinear in comparison to quasi steady‐state swimming in recirculating flumes, they were much less unsteady than the volitional swimming behaviors that underlie existing predictive models of field swimming cost. Performance analyses suggested that speed and path curvature data could be used to derive reasonable estimates of locomotor cost that fit within measured capacities for sustainable activity. The distinct differences between field swimming behavior and performance measures obtained under steady‐state laboratory conditions suggest that field observations are essential for informing approaches to quantifying locomotor performance in the laboratory.