It is a clinical impression that less fentanyl is needed for anesthesia during hyperventilation and hypocarbia. If true, it might be due to both increased penetration of fentanyl, a highly lipid-soluble agent, into the brain and increased brain tissue binding. Serum and brain concentrations of fentanyl were determined in dogs anesthetized with halothane during normocarbia, hypocarbia by hyperventilation, and hypercarbia by addition of CO2 to the inspired mixture. Fentanyl, 12.5 μg/kg, was injected iv, and serum and brain samples were taken for fentanyl analysis by radioimmunoassay. Brain fentanyl values peaked latest (15–20 min) and were highest during hypocarbia; brain fentanyl values peaked earliest (0–5 min) and were lowest during hypercarbia; values during normocarbia were intermediate in time to peak (10–15 min) and concentration. Thereafter, brain levels declined, but during hypocarbia were significantly higher and during hypercarbia were significantly lower than during normocarbia. Interestingly, serum fentanyl levels were also significantly higher during hypocarbia. The brain-blood fentanyl ratios for each of the three CO2 levels increased for 30 min and thereafter stayed relatively constant. The brain-blood ratios were highest with hypocaria and lowest with hypercarbia. At 35 min, when clinical analgesia may be considered termineated, hypocarbic brain levels were double those of normocarbia. The authors feel this reflects, to a large extent, higher serum fentanyl concentrations and delayed cerebral wash-out because of decreased blood flow. To a small but unknown extent the higher brain fentanyl levels result from increased brain-blood penetration due to increased lipid solubility, and increased brain tissue binding of fentanyl during respiratory alkalosis.