Although clinical rating scales and simple timed tests of motor function are widely used to assess motor response to therapy, gait analysis may provide an alternative measure of this response. We studied 15 patients with PD complicated by motor fluctuations, first to determine changes in temporal and spatial gait parameters following levodopa, secondly to assess the stability of repeated gait measures and timed tests in "off" and "on" states, and thirdly to determine the use of gait analysis in the assessment of the dopaminergic response. Gait analysis (velocity, stride length, cadence, and double limb support), clinical rating scales (modified Webster scale and Hoehn and Yahr stage), and timed tests of motor function (hand tapping and stand-walk-sit time) were performed before ("off") and after ("on") a levodopa challenge. Stride length and gait velocity increased following medication whereas cadence and double limb support did not. Most gait measures and the stand-walk-sit time were stable over three consecutive trials in both "off" and "on" states. Of the gait measures, only cadence in the "off" state changed significantly whereas the tapping count improved with repeated trials in both "off" and "on" states. Changes in stride length, gait velocity, and tapping count following levodopa correlated with changes in clinical rating scales following treatment. Measurement of gait parameters provides a reliable, objective alternative to rating scales and timed tests in assessing the dopaminergic response in patients with PD and motor fluctuations.