Dog models of total brain ischaemia for the study of pathophysiological mechanisms and post-resuscitative therapy used since 1970 are reviewed. Attempts to produce isolated total cerebral ischaemia in dogs either failed to stop cerebral circulation completely because of the abundance of collateral vessels or involved complex operations with blood loss. Complete airway obstruction resulted in various periods of asphyxia preceding asystole. Total circulatory arrest by ventricular fibrillation for 12 min or longer, resulting in gross neurological deficit, required cardiopulmonary resuscitation of variable duration and often required thoracotomy to restart spontaneous circulation. An acceptable model for short-term studies was obtained by clamping the ascending aorta for 15 min. This resulted in control of perfusion pressure with precise timing of onset and termination of cerebral ischaemia, and an isoelectric electroencephalogram within a mean time of 37 (26-44) s; it retained its ability to restore spontaneous circulation promptly with special supportive therapy, and a severe neurological deficit was evident for 20 h after ischaemia. Reliable total brain ischaemia suitable for long-term survival studies could not be developed in the dog without major surgical procedures, as it has been in the rhesus monkey. Sixteen minutes of high-pressure neck occlusion with a tourniquet and special intensive care resulted in major neurological deficit with survival for at least 7 days, and proved suitable for simultaneous study of neurological, physiological and histological changes.