BACKGROUND: Vitamin D may improve muscle strength through a highly specific nuclear receptor in muscle tissue. OBJECTIVES: We investigated whether there is an association between 25-hydroxyvitamin D [25(OH)D] concentrations and lower-extremity function in ambulatory older persons, whether that association differs by activity level, and, if so, whether there is an identifiable threshold in the association. DESIGN: The study was a population-based survey of the ambulatory US population aged 60 to > or =90 y (n = 4100). Lower-extremity function according to serum 25(OH)D concentrations was assessed by linear regression analyses and regression plots after control for activity level (inactive or active) and several other potential confounders. Separate analyses were performed for the timed 8-foot (ie, 2.4 m) walk test and a repeated sit-to-stand test. RESULTS: The 8-foot walk test compared subjects in the lowest and highest quintiles of 25(OH)D; the latter group had an average decrease of 0.27 s [95% CI: -0.44, -0.09 s (or 5.6%); P for trend < 0.001]. The sit-to-stand test compared subjects in the lowest and highest quintiles of 25(OH)D; the latter group had an average decrease of 0.67 s [95% CI: -1.11, -0.23 s (or 3.9%); P for trend = 0.017]. In the 25(OH)D reference range of 22.5-94 nmol/L, most of the improvement occurred in subjects with 25(OH)D concentrations between 22.5 and approximately 40 nmol/L, and further improvement was seen in the range of 40-94 nmol/L. Stratification by activity level showed no significant effect modification. CONCLUSION: In both active and inactive ambulatory persons aged > or =60 y, 25(OH)D concentrations between 40 and 94 nmol/L are associated with better musculoskeletal function in the lower extremities than are concentrations < 40 nmol/L.