Sixteen acceptably randomized studies of anticoagulant therapy after cerebral or retinal ischemia or infarction are reviewed and the results among 1,046 anticoagulated patients and 1,071 controls are analyzed. The following conclusions are derived. 1) Anticoagulant therapy has not been shown to be better than control management after transient ischemia or nonprogressing ischemic stroke; this is true whether the control management was deliberately ineffectual treatment (generally studies completed in 1974 or earlier) or platelet antiaggregant therapy (pooled results of three recent studies). 2) Although a study done 30 years ago demonstrated no benefit, a recent study showed benefit from anticoagulant therapy in patients who had had cerebral emboli of cardiac origin; additional controlled data are needed. 3) There is evidence that patients with thrombosis in evolution might benefit from anticoagulant therapy; additional controlled data are needed.