Two hundred eight witnessed subjects dying suddenly (SD) (151 in 10 minutes, 47 in one hour, and 10 in three hours) and 97 dying by accident (AD) were studied. All these subjects had no clinical history of CHD. They were at the time of death in their normal, usual activity, and not under medical care. No therapy before and resuscitation attempts after death were done. The unique postmortem findings were coronary atherosclerosis, and/or myocardial necrosis or fibrosis. Coronary stenosis was absent or insignificant (< 50 per cent) in 28, median in 23, and severe (≥ 70 per cent) in 157 (53 monovessel, 60 double vessel, 44 triple vessel or more) SD cases. In AD the figures were 28, 31, 38 (26, nine, and three), respectively. Thirty-two (15.3 per cent) SD subjects showed an acute occlusive thrombus located in a severe, long, concentric stenosis. An infarct was demonstrable in 35 SD cases (11 plus occlusive thrombus). A necrosis resembling that induced by catecholamines was the unique acute myocardial lesion found in 149 SD subjects (72 per cent). This lesion was also present in 29 out of 35 subjects with an infarct (total 178 = 85.5 per cent), in 22 out of 28 SD cases with insignificant coronary damage, and in 19 AD cases. One hundred two SD cases with prodroma had a significant lower frequency of thrombi and a higher frequency of stenosis ≥ 90 per cent, triple vessel involvement, myocardial fibrosis, and pathologic heart weight (≥ 500 gm.). These data suggest that: (1) there is no proof that SD is due to a morphologic occlusive cause, the thrombus being likely to be an ineffective event secondary to regional increased peripheral resistance in the presence of functioning collaterals; (2) there is no linear cause-effect relation between the degree of coronary damage and SD; (3) a primitive sympathetic disorder may be responsible for ventricular fibrillation, myocardial necrosis/fibrosis, and cardiac hypertrophy-SD being not necessarily synonymous with infarct. © 1979.