A combined approach utilizing ophthalmosonometry (OSM), carotid phonoangiography (CPA), and ocular pneumoplethysmography (OPG) was applied to 31 patients with symptoms of carotid artery occlusive disease. Arteriograms were subsequently obtained in 70 of these patients, thus allowing the accuracy of each technique to be assessed in 140 carotid arteries. Both Doppler OSM and OPG correctly identified all complete occlusions of the internal carotid artery. However, greater than 50% stenoses were detected by OSM in only 52% and by OPG in 87%. Thus the cumulative accuracy in detecting hemodynamically significant lesions was 76% for OSM and 93% for OPG. The addition of CPA to OPG raised the combined accuracy of the two techniques to 98% for significant carotid lesions. The development of a regression line criterion for the OPG, in combination with CPA, permitted identification of 100% of patients with bilateral carotid artery lesions. However 75% of patients with symptomatic, ulcerating plaques were missed by all three tests, thereby emphasizing the need for angiography in appropriately symptomatic patients. The principal usefulness of noninvasive testing in evaluating patients with carotid artery occlusive disease is for: (1) asymptomatic bruits, (2) atypical, nonhemispheric symptoms, or (3) following carotid endarterectomy.