Thirty-six institutionalized geriatric patients with particularly severe and longstanding dementia were selected by chart review and were studied neurologically. Criteria for clinical diagnosis are derived, and it is shown that arteriosclerosis had caused or contributed to their dementias in 33 patients (91.7%) and probably did so in two others (97.2%). Arteriosclerotic dementia may not be rare and may be much more prevalent than may now be considered the case. As 23 patients (63.9%) showed evidence of carotid and/or vertebrobasilar involvement, large-vessel disease also may be more prevalent in geriatric dementia than is now thought and should be considered in all such patients. The results further confirm that careful review of ward records without reference to psychiatric diagnoses of etiology is a simple, rapid, and reliable method of screening a large institutionalized population for possible arteriosclerotic dementia. The selectees should then be surveyed appropriately to confirm the presence of the disease. Since a careful review of the complete hospital records and a thorough neurologic examination proved sufficient, more definitive but more dangerous invasive diagnostic procedures may not be necessary for clinical pharmacologic and similar investigations.