Within a group of 2,000 patients evaluated, most of them with symptoms of cerebrovascular insufficiency, 441 had a monolateral or bilateral cervical bruit. The 627 sides with an audible bruit were divided into main groups (A) symptomatic (TIA and/or stroke homolateral to the bruit), (B) possibly symptomatic (non-side-related symptoms), (C) asymptomatic (C1, in totally asymptomatic patients; C2, in patients with symptoms dependent on the hemisphere contralateral to the cervical bruit). Each patient was studied by means of clinical (history, blood pressure in both upper limbs, phonoendoscopic auscultation at various levels) and noninvasive instrumental examinations (CW Doppler spectrum analysis). An apparently primitive cervical bruit corresponded to a lesion of the carotid bifurcation in 61% of the cases (positive predictive value) whereas a normal bifurcation was detected in 70% of the cases in which the cervical bruit was considered as secondary (negative predictive value); the diagnostic accuracy of the "critical auscultation" has a value therefore of 63%, with a sensitivity of 84% and a specificity of 40%. The results obtained in the different groups of patients (symptomatic or asymptomatic) were not significantly different (chi-square). Even though maintaining the value of a cervical bruit as a sign of carotid stenosis or occlusion and consequently confirming the importance of neck auscultation, the authors conclude that the critical auscultation as commonly performed is not capable of excluding the presence of a carotid lesion with sufficient reliability, even in totally asymptomatic patients.