A case-control study of Parkinson's disease and physical exercise was conducted in a cohort of 50,002 men who attended Harvard College (Cambridge, Mass) or the University of Pennsylvania (Philadelphia) between 1916 and 1950 and were followed up in adulthood for morbidity and mortality data. Cases of Parkinson's disease were identified from responses to mailed questionnaires and death certificates through 1978. Four controls from the same population were selected for each case. The association between physical activity at the time of college and subsequent risk of Parkinson's disease was evaluated for 137 cases and 548 controls, whereas the data on physical activity in adulthood before the disease occurrence was available only for 94 of these cases. Having belonged to a varsity team or having done regular physical exercise in college was associated with a lower nonsignificant risk of Parkinson's disease. In adulthood, practice of moderate or heavy sports was linked to a reduced risk, although more precise analysis revealed that there was only a modest nonsignificant reduction in risk for subjects who do a moderate amount of physical exercise, but this negative association disappears at higher levels of physical expenditure. These results, which require further confirmation, are compatible with a slight protective effect of physical exercise on the risk of Parkinson's disease, although the lack of association cannot be refuted.