In 632 consecutive primary operations for coronary bypass grafting, the effect of instrumentation of the ascending aorta on the prevalence of stroke was evaluated. There were five surgical pathology groups: group A: soft aorta with no palpable disease, 463 patients; group B: distinctly palpable aorta with focal atheromas necessitating minor surgical modifications, 132 patients; group C, unclampable aorta (no plane for crossclamping the aorta without compression of atheromas was present), 16 patients; group D, untouchable aorta, in which the entire ascending aortic wall was involved by atheromatosis (these aortas were not touched), 14 patients; group E, aneurysmal aorta with soft walls, 7 patients. There were four strokes (0.63%), all related to instrumentations of the aorta. There were no cerebrovascular accidents in patients in whom precautions were taken. Of all risk factors studied, age was the only predictive one for major atheromatosis of the aorta (chi 2 test, p less than 0.001). Of the 30 patients in groups C and D, only one was younger than 60 years.