The functions of fish skin during swimming remain enigmatic. Does skin stiffen the body and alter the propagation of the axial undulatory wave? To address this question, we measured the skin's in situ flexural stiffness and in vivo mechanical role in the longnose gar Lepisosteus osseus. To measure flexural stiffness, dead gar were gripped and bent in a device that measured applied bending moment (N m) and the resulting midline curvature (m-1). From these values, the flexural stiffness of the body (EI in N m2) was calculated before and after sequential alterations of skin structure. Cutting of the dermis between two caudal scale rows significantly reduced the flexural stiffness of the body and increased the neutral zone of curvature, a region of bending without detectable stiffness. Neither bending property was significantly altered by the removal of a caudal scale row. These alterations in skin structure were also made in live gar and the kinematics of steady swimming was measured before and after each treatment. Cutting of the dermis between two caudal scale rows, performed under anesthesia, changed the swimming kinematics of the fish: tailbeat frequency (Hz) and propulsive wave speed (body lengths per second, L s-1) decreased, while the depth (in L) of the trailing edge of the tail increased. The decreases in tailbeat frequency and wave speed are consistent with predictions of the theory of forced, harmonic vibrations; wave speed, if equated with resonance frequency, is proportional to the square root of a structure's stiffness. While it did not significantly reduce the body's flexural stiffness, surgical removal of a caudal scale row resulted in increased tailbeat amplitude and the relative total hydrodynamic power. In an attempt to understand the specific function of the scale row, we propose a model in which a scale row resists medio-lateral force applied by a single myomere, thus functioning to enhance mechanical advantage for bending. Finally, surgical removal of a precaudal scale row did not significantly alter any of the kinematic variables. This lack of effect is associated with a lower midline curvature of the precaudal region during swimming compared with that of the caudal region. Overall, these results demonstrate a causal relationship between skin, the passive flexural stiffness it imparts to the body and the influence of body stiffness on the undulatory wave speed and cycle frequency at which gar choose to swim.