NMR in vivo spectroscopy is one of the few methods available for non-invasive investigations of cerebral metabolism in animals and humans. 31P and 1H spectroscopy are particularly suitable for monitoring the cerebral energy metabolism by determining the cerebral levels of ATP, ADP, phosphocreatine (PCr), inorganic phosphate (Pi), lactate and intracellular pH (pHi). These techniques also seem to be suitable for studying the effects of anesthetics by directly comparing the anesthetized and unanesthetized states in the same subject. The effects of halothane and isoflurane on the changes elicited in the cerebral energy metabolism by experimental hypercapnia were investigated by in vivo NMR spectroscopy. Halothane was found to aggravate the decrease in PCr attributed to the shift in creatine kinase equilibrium induced by the cerebral acidosis associated to hypercapnia, while the level of cerebral ADP was decreased to a lesser extent than in unanesthetized animals. In contrast isoflurane did not modify the changes in cerebral energy metabolism elicited by hypercapnia except that the decrease in PCr was significantly slowed, suggesting a lower creatine kinase activity. These data indicate that isoflurane and halothane act by two different mechanisms to produce a decrease in oxygen consumption. Halothane could interfere with oxidative metabolism by disturbing ATP metabolism, while isoflurane could decrease oxygen consumption by a general sedative action, slowing both cerebral functional activity and cerebral energy homeostasis.