Patients with Parkinson's disease treated with levodopa over the past 15 years were compared, by parallel statistical methods, to a group of similar patients followed for 15 years before the levodopa era. The duration of illness at each Stage of disease was longer in the treated group, and, at each duration of illness, there was less disability and death. The prevalence of peak-dose dyskinesias increased with increased duration of treatment, but seldom out-weighed the benefits of treatment. Although extreme fluctuations of therapeutic response were not seen during the first 2 years of treatment, their prevalence was otherwise uninfluenced either by the duration of treatment or by postponing treatment. There is, however, some evidence that the postponement of treatment is accompanied by an increased proportion of patients who become "unresponsive" to levodopa. The age at death was 4.5 years older than in the untreated group, and the mortality rate equal to that of the general population. There is ample evidence that treatment with levodopa improves the quality and length of life, and no real evidence that delaying therapy confers benefits in the future.