We examined 15 patients with cortical blindness, reviewed the records of 10 others, and compared these 25 patients to those in previous studies of cortical blindness. Although cerebrovascular disease was the most common cause in our series, surgery, particularly cardiac surgery, and cerebral angiography were also major causes. Only 3 patients denied their blindness, although 4 others were unaware of their visual loss. Electroencephalograms (EEGs) were performed during the period of blindness in 20 patients and all recordings were abnormal, with absent alpha rhythm. Visual evoked potentials recorded during blindness were abnormal in 15 of 19 patients, but did not correlate with the severity of visual loss or with outcome. Bioccipital lucencies were found in computed tomographic (CT) scans of 14 patients; none of the 14 regained good vision. Recovery of vision was poor in all 8 patients who had a spontaneous stroke, but fair or good in 11 of the other 17 patients. Prognosis was best in patients under the age of 40 years, in those without a history of hypertension or diabetes mellitus, and in those without associated cognitive, language, or memory impairments. We conclude that the prognosis in cortical blindness is poor when caused by stroke; EEGs are more useful than visual evoked potentials for diagnosis; and bioccipital abnormalities shown on CT scan are associated with a poor prognosis.