This study examined whether inability to perform adaptive locomotor tests predicts self-reported incident mobility disability. InCHIANTI study participants (N = 611; age, 50-85 yrs) who could walk 7 m at self-selected speed and who had no self-reported mobility disability at baseline were included. The ability to complete four adaptive locomotor tests was assessed: fast walking, walking on a narrow path, crossing obstacles while walking, and talking while walking. Mobility disability was recorded again at 3-yr follow-up. Failure in the fast-walking and narrow-path walking tests predicted approximately 2.5 times likelihood of reporting incident mobility disability (P = 0.009 and P = 0.011, respectively). Failure in the obstacle-crossing test predicted approximately two times likelihood of reporting incident mobility disability; however, this result did not reach statistical significance (P = 0.077). Failure in talking while walking did not predict incident mobility disability. Those who failed both the fast-walking and narrow-path walking tests were almost nine times as likely to report incident mobility disability.