To determine the effects of the stooped posture of patients with Parkinson's disease (PD) on postural stability, we compared the kinetic, kinematic, and electromyographic responses of seven subjects with PD and 11 control subjects to eight directions of surface translations. Control subjects were studied in an upright posture and in a stooped posture that mimicked the posture of the PD subjects. When control subjects adopted a stooped posture, peak center of pressure displacements slowed and decreased, reducing stability margins toward values observed in PD subjects. Stooped control subjects, however, responded to translations with large joint angle displacements, whereas PD subjects exhibited small joint angle displacements. Stooping in control subjects did not lead to abnormally directed horizontal forces under each foot or antagonistic muscle co-activation at the hip and trunk, as seen in PD subjects. Upright and stooped control subjects never fell during the trials, whereas PD subjects fell in 16% of the trials. We conclude that stooped posture is a destabilizing posture, but it does not account for abnormal postural responses in PD.