170 patients with suspected temporal arteritis underwent Doppler investigation before temporal artery biopsy. Doppler study included the large cervical trunks, the ophthalmic system and analysis of the terminal branches of the external carotid with recording of the temporal territory on 3 points. Biopsy confirmed the diagnosis of temporal arteritis in 48 cases; it was negative in 122 cases. Doppler study and histological examination agreed in 137 patients (80%). There was disagreement in 33 cases (20%). The Doppler signal was pathological in 43 of the 48 cases of temporal arteritis (90%). There were 5 false negatives (10%). In subjects free from temporal arteritis, the Doppler examination was normal in 94 cases (77%), and was pathological in 28 cases (23%). The sensitivity of Doppler examination in the diagnosis of temporal arteritis was 90%, its specificity was 77%. The predictive value of a positive Doppler vis-a-vis the condition was only 61%. On the other hand, the predictive value of a normal examination was very high (95%). The high level of false positives is explained by technical problems, but especially by a high incidence of sometimes very stenosing senile arteritis. Finally, 3 of the 28 false positives (11%) subsequently developed authentic inflammatory arteritis. The authors conclude that a positive Doppler examination does not allow reliable prediction of the presence of temporal arteritis. On the other hand, a normal examination almost always allows the diagnosis of temporal arteritis to be excluded with a risk of error of less than 5%.