The authors investigated performance in 2 rhythm tasks in young (M = 23.8 years) and older (M = 71.4 years) amateur pianists to test whether slowing of a central clock can explain age-related changes in timing variability. Successive keystrokes in the rhythm tasks were separated by either identical (isochronous) time intervals or varying (anisochronous) intervals. Variability was comparable for young and older adults in the isochronous task; pronounced age effects were found for the anisochronous rhythm. Analyses of covariances between intervals rule out slowing of a central clock as an explanation of the findings, which instead support the distinction between target specification, timekeeper execution, and motor implementation proposed by the rhythm program hypothesis (D. Vorberg & A. M. Wing, 1996). Age stability was found at the level of motor implementation, but there were a.-c-related deficits for processes related to target-duration specification.