Sixty consecutive patients with visual complaints as the primary symptom of cerebrovascular insufficiency were evaluated with complete cerebral angiography (CCA). Eighteen patients referred for evaluation of asymptomatic Hollenhorst plaques had a 50% incidence of ipsilateral carotid bifurcation disease, but only 16% of these were considered surgically remediable; five of these patients eventually required carotid endarterectomy (CEA). Nine of the 18 patients with asymptomatic Hollenhorst plaques were further evaluated with serial ocular fundus photography: the plaque persisted in all nine patients--in four for more than 6 months and in five for more than 1 year. No patient had transient or fixed retinal or cerebral symptoms. Twenty-six patients with amaurosis fugax showed a 70% incidence of significant bifurcation disease and three had significant intracranial disease; 50% eventually required CEA. Twelve patients with fixed visual deficits (four patients had homonymous hemianopsia, four had retinal artery occlusions, and four had ischemic retinopathies) demonstrated a 100% incidence of bifurcation disease, but in addition 85% had significant intracranial disease; only one of this group had surgically amenable disease. Amaurosis fugax preceded blindness in two patients. Hollenhorst plaque as an isolated phenomenon appears to be a retinal embolic event of uncertain age, with poor predictive power for further embolic events, and not as highly correlated with significant or surgically amenabe carotid bifurcation disease.