Patients with proliferative diabetic retinopathy had one eye chosen by a randomization procedure to have treatment by xenon-arc photocoagulation. 100 patients were followed for at least 1 year, 58 patients for 2 years, and 23 for 3 years. The patients form roughly two equal subgroups: those with new vessels on both optic discs and those without new vessels on either disc. There was a significant difference between the visual acuity of the treated and the untreated eyes in the former group, i.e. with disc new vessels, after 1, 2, and 3 years, but there was no difference in the latter group. Eleven patients were blind in one or both eyes for two consecutive yearly assessments. No treated eye became blind without concomitant blindness in the fellow eye but eight untreated eyes did so. This difference was statistically significant.