The final results of a randomised controlled study of xenon-arc photocoagulation for proliferative retinopathy are reported, after all patients have been followed for at least 5 years and some for up to 7 years. One hundred and seven patients with two similarly affected eyes had one treated (chosen by a random procedure), while the other eye remained untreated and served as a control. Of the 107 patients, 77 completed the 5 year follow up, 13 died and 17 stopped attending for various reasons. Of the recorded coexistent medical abnormalities, only renal complications affected survival, none influenced visual outcome. Visual outcome was significantly better in the treated than in the control eyes at each yearly interval (0.001 less than p less than 0.05). The greatest difference was seen in those with disc vessels at entry. In this group, control eyes deteriorated by a mean of four lines on the Snellen chart, treated eyes by one line only. Six patients became legally blind in both eyes, four were blind in the treated eye only, but 28 control eyes were blind when treated eyes retained vision (p less than 0.001). Treated eyes which became blind had less treatment than those that retained vision. Of the 42 treated eyes with peripheral new vessels only at entry, 12 developed disc new vessels. These 12 had fewer burns than the 30 which did not develop disc new vessels. It is concluded that in proliferative retinopathy, treatment by photocoagulation is better than no treatment at all. Adequate treatment is required to maintain vision.