This pilot study was performed to compare the occurrence of long term motor complications in Parkinson's disease when the introduction of levodopa was delayed by an initial treatment with high doses of bromocriptine alone. The trial was a prospective randomised controlled study comparing 31 previously untreated patients with Parkinson's disease initially given bromocriptine alone to which levodopa was later added (group B/D) and 29 other previously untreated patients with Parkinson's disease immediately given levodopa alone (group D). The end point was the occurrence of the first motor complications (wearing off or dyskinesia). Group B/D patients received bromocriptine (52 (SEM 5) mg/day) for 2.7 years, to which levodopa was later added (471 (SEM 46) mg/day). Group D patients received a comparable dose of levodopa alone (569 (SEM 47) mg/day). Both had similar disability scores at the end of the study. Motor complications were fewer and appeared later in group B/D than in group D (56% after 4.9 (SEM 0.5) years of treatment v 90% after 2.7 (SEM 0.5) years, p < 0.01). Wearing off appeared later (p < 0.01) in group B/D (4.5 (SEM 0.6) years) than in group D (2.9 (SEM 0.6) years). Peak dose dyskinesia occurred less often in group B/D patients (three v 14 cases, p < 0.01). This study showed that a three year initial monotherapy with high doses of bromocriptine followed by addition of levodopa delayed the occurrence of long term motor complications usually found in patients with Parkinson's disease treated with levodopa alone from the beginning.