Subsurface, aquatic locomotion has evolved independently at least ten times across the avian phylogeny. Given the unique challenges associated with aquatic locomotion, it is uncertain how the re-invasion of the aquatic environment could occur. Water is 800 times denser and 60 times more viscous than air. Thus, secondary adaptation to water presumably requires significant morphological and physiology changes. One option is that lineages are ‘pre-adapted’ to subsurface aquatic locomotion through selection for efficient locomotion on the surface of the water. In our efforts to study the similarities between surface-based aquatic locomotion in non-aquatic birds and the subsurface swimming of semi-aquatic birds, we made a surprising discovery. European starlings (Sturnus vulgaris), a non-aquatic species, gracefully swim when placed underwater. Though this work is in its early stages, we can report a few observations relevant for understanding the evolution of subsurface swimming in birds. First, starlings use both their feet and wings for aquatic locomotion, but the feet appear to function primarily for directional control, as in true wing-propelled diving birds. Second, both the upstroke and downstroke of the wings are hydrodynamically active, with the upstroke producing negative heave and the downstroke producing both positive heave and thrust. And third, starlings adopt the flexed-wing posture exhibited by all volant, wing-propelled diving birds, suggesting that this posture is the result of biomechanical constraints rather than selection for efficient swimming. We plan to extend this work to more non-aquatic species, test whether individuals improve their swimming performance through learning, and rigorously compare the 3D kinematics of swimming in non-aquatic species to that of true semi-aquatic birds, including the European starling’s close relative, the American dipper.