Plasma levodopa and therapeutic responses to treatment with levodopa in combination with benserazide or carbidopa were studied in 49 patients with Parkinson disease not previously treated with levodopa in a blind randomized crossover trial. The treatment periods were 12 weeks; similar dosage schedules were used, with doses that induced equal levels of plasma levodopa in both combinations. In pretrial studies of plasma levodopa responses, 200 mg of levodopa and 50 mg of benserazide was equal to 250 mg of levodopa combined with 25 mg of carbidopa. Equal plasma levodopa responses to both combinations were also found during the trial. There was no significant difference between the treatment groups in beneficial effects on parkinsonian disability and individual symptoms or in the frequency of involuntary movements. However, nausea and vomiting occurred significantly more often during treatment with levodopa and carbidopa than during treatment with levodopa and benserazide. This difference was probably due to inadequate inhibition of peripheral decarboxylase inhibitor by the 1:10 ratio of carbidopa to levodopa.