The authors used in vivo 19F nuclear magnetic resonance spectroscopy to determine rates of cerebral uptake and elimination of desflurane, isoflurane, and halothane in rabbits. After anesthetizing animals by intramuscular and intravenous injection of methohexital and inhalation of 70% nitrous oxide, intravenous and intraarterial catheters were inserted and a tracheostomy and craniotomy performed. Ventilation was controlled to maintain arterial carbon dioxide tension (PaCO2) from between 35 and 45 mmHg. A 2-2.5-cm diameter circle of dura was exposed, over which a 0.9 x 1.0-cm elliptical surface coil was placed. Cerebral anesthetic concentrations (CC) were estimated from spectra acquired on a 4.7-Tesla spectrometer. Alveolar uptake and elimination also were assessed, using inspired (FI) and end-tidal (denoted FA0 at the end of administration) concentrations measured by gas chromatography. After baseline spectra were obtained, volatile agents were administered for 30 min, followed by a 120-min period of elimination. Our findings demonstrate that cerebral uptake and elimination correlate with solubility: they are most rapid for desflurane, next most rapid for isoflurane, and least rapid for halothane. During administration, cerebral uptake of desflurane (CC/FI = 0.690 +/- 0.049 at 9 min) was approximately 1.7 times faster than isoflurane (CC/FI = 0.691 +/- 0.020 at 15 min) and 3 times faster than halothane (CC/FI = 0.662 +/- 0.040 at 27 min). Similarly, elimination rates for desflurane (CC/FA0 = 0.238 +/- 0.015 at 9 min) were 1.7 times faster than isoflurane (CC/FA0 = 0.236 +/- 0.017 at 15 min) and three times faster than halothane (CC/FA0 = 0.212 +/- 0.033 at 27 min).(ABSTRACT TRUNCATED AT 250 WORDS)