Four geographically diverse centers provided data on mortality in 359 patients with Parkinson's disease, the majority of whom began dopa treatment during the early experimental trials of 1968 to 1970. Patients were classified into three groups based on the duration of symptoms prior to starting dopa treatment: Group 1, 1 to 3 years; Group 2, 4 to 6 years; Group 3, 7 to 9 years. After 15 years of treatment and 3,689 person-years of observation, Group 1 had an observed-to-expected mortality ratio of 1.43; Group 2, 2.44; and Group 3, 2.95 (p less than 0.05). This result confirmed that increased duration of disease was associated with increased mortality risk. To examine the effect of the time of initiation of dopa treatment, duration of disease was held constant at 17 years for all three groups. Observed-to-expected mortality ratios were 1.43 for Group 1; 2.66 for Group 2; 2.63 for Group 3. This statistically significant advantage for Group 1 (p less than 0.0001) led to the conclusion that early treatment with dopa has a beneficial effect on life expectancy. After 17 years of disease, causes of death in Group 1 were less likely (p = 0.027) to be due to Parkinson's disease than was found in the other groups.