1. Percutaneous transluminal dilatation was performed in 37 patients with renovascular hypertension: 24 had atherosclerotic renal artery stenosis and 13 had fibromuscular dysplasia. 2. Percutaneous transluminal dilatation could not be performed for technical reasons in three (8%) of the 37 patients. 3. In the remaining 34 patients blood pressure fell significantly (P < 0.001) from 201 ± 31/118 ± 14 mmHg to 144 ± 22/91 ± 11 mmHg 3 days after the procedure. The respective values at months 6 and 24 were 148 ± 26/89 ± 12 mmHg (n = 23, P < 0.001) and 143 ± 14/89 ± 6 mmHg (n = 8, P < 0.001). 4. Certain differences between the two subgroups emerged in the response to percutaneous transluminal dilatation (6 months values, n = 23): cure rate was higher in patients with fibromuscular dysplasia than in those with atherosclerotic stenosis (67% vs 35%) and in contrast to atherosclerotic stenosis none of the cases with fibromuscular dysplasia was unimproved. 5. Follow-up angiography performed at month 6 showed recurrence of slight renal artery stenosis in five out of 19 patients (all atherosclerotic). 6. Complications were seen in six (16%) of the 37 patients. 7. Our results show that percutaneous transluminal dilatation is a valid method for the treatment of renovascular hypertension. Patients with fibromuscular dysplasia may show a better response than those with atherosclerotic stenosis. In this study the latter was complicated by a high risk of developing restenosis. Finally a complication rate of 16% requires careful selection of patients.