Transcranial Doppler sonography (TCD) has gained in relevance for noninvasive monitoring of the cerebral circulation during the perioperative period. As long as the diameters of the investigated vessels remain unknown, however, flow velocities alone are not really informative. Exact vessel diameter determination in humans under the influence of different anesthetic drugs has not yet been performed due to ethical and methodological restrictions. A new modification of TCD allows analysis of the reflected "Doppler power", which is proportional to the cross-sectional area of the insonated vessel. METHODS. Three groups of 15-16 patients each (ASA I) were investigated after informed consent. Anesthesia was induced with droperidol, alfentanil, thiopental, and vecuronium bromide. After intubation patients were normoventilated with N2O:O2 = 3:2 and additional doses of alfentanil were injected until the transcranial ultrasound probe was fixed to the temporal bone and focused on the middle cerebral artery. Baseline values of heart rate (HR), mean arterial pressure (MAP), expiratory minute volume (EMV), end-expiratory CO2 (FeCO2), and TCD were measured. Then 1.5 vol% halothane, 25-50 micrograms/kg alfentanil, or propofol (1.5 mg/kg iv., 10 mg/kg.min) was administered. Further measurements (HR, MAP, EMV, FeCO2 and TCD) were performed at 3, 6, 10, and 20 min after the start in the halothane and propofol groups and after 3 and 6 min in the alfentanil group. The following transcranial parameters were derived from the TCD device: mean maximal flow velocity (vmmax), pulsatility index, time-averaged mean velocity (vmmen), "vessel area (VA)", and "volume flow (VF)". The mean +/- standard deviation of each parameter was calculated. Statistical evaluation was performed by paired t-tests (level of significance P less than 0.05). RESULTS. HR showed a tendency to increase after halothane and to decline after alfentanil. Alfentanil induced a short-term decline in MAP. End-expiratory minute volume and FeCO2 showed only minor alterations in all three groups. The vmmax was nearly doubled by halothane. Alfentanil induced a transitory decline in vmmax while Propofol decreased it markedly. The pulsatility index showed a decrease after halothane; alfentanil caused a short-term increase. Propofol induced a strong increase after 3 min; in the following period a return to baseline values was observed. The vmmean was increased by halothane and reduced by 32% propofol. VA was found to be unaltered by alfentanil and propofol but was more than doubled by halothane. Accordingly, the relative value for VF increased by 148% under halothane. VF appeared to decline after propofol. DISCUSSION. The described method allows only the determination of relative values: it is not possible to quantify exactly how much the VA changed. Halothane caused significant increases of VA measured in the middle cerebral artery, whereas alfentanil and propofol did not influence this parameter. This is in accordance with previous experiments in dogs in which halothane decreased the resistance of large basal cerebral arteries (LAR). LAR remained unaltered after alfentanil administration. The site of action of some anesthetic agents on cerebral vessels does not seem to be restricted to cerebral arterioles: at least for halothane, a vasodilating effect on large cerebral arteries could be demonstrated. This should be kept in mind when transcranial Doppler is applied during the perioperative period.