Clinical scales, based on the major signs of Parkinson's disease (PD), are commonly used to assess the effect of symptomatic treatment of PD. With the appearance of therapy aiming to rescue or protect the nigrostriatal neurons in PD, it becomes essential to define which of these signs best reflects the underlying neuronal deficit. Fluorodopa positron emission tomography has been shown to correlate with postmortem nigral cell counts. We correlated the major signs of PD with positron emission tomography results in 35 PD patients. We found that in the "practically defined off" state, (1) Purdue pegboard scores correlated best with the nigrostriatal dopaminergic deficit; (2) of the subscales of the modified Columbia score, the bradykinesia subscale correlated best; (3) rigidity and postural disturbance correlated less highly than bradykinesia, and their inclusion in a multiple regression did not improve the correlation of pegboard or bradykinesia scores alone; and (4) tremor did not correlate with the nigrostriatal dopaminergic deficit. We conclude that pegboard and bradykinesia scores represent the best clinical measures for studying the effect of treatment on the evolution of the nigrostriatal lesion of PD. Inclusion of other clinical signs provides additional information only for the study of functional impairments.