We used four disability scales to evaluate eight patients with Parkinson's disease who were treated with pergolide mesylate for 1 year. Disability was rated on all four scales by the same neurologist at each of 11 visits. Prior ratings were not available to the examiner, who did not know that the scales themselves were an object of study. Disability scores, converted to percentage improvement relative to baseline, varied considerably between scales; for instance, at 5 months, one showed 13% improvement and another 58%. At 9 months, one showed worsening of 6% and another showed improvement of 34%. The four disability scales clearly measure different aspects of parkinsonism, and comparing results of different studies may not be valid if the disability scales are not the same.