Tissue oxygen (PO2), K+ (aKe), pH (pHe) and Ca2+ ([Ca2+]e) were measured in the region of the ventral respiratory group (VRG) in the in vitro brainstem-spinal cord preparation of neonatal rats. During tissue anoxia, elicited by superfusion of N2-gassed solutions, an initial increase in the frequency of respiratory activity, lasting between 2 and 12 min, turned into a frequency depression. During anoxia periods of up to 60 min, respiratory activity persisted in solutions containing CO2/bicarbonate, whereas a complete blockade was observed after 15-25 min in N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid- (Hepes)-buffered salines. After such anoxic apnea, respiratory rhythmicity could be reactivated by superfusion of hypoxic, CO2/bicarbonate-buffered solutions. In both types of hypoxic solutions, aKe increased by maximally 1.5 mM, whereas an initial increase of pHe by up to 0.05 pH units turned, after 2-4 min, into an acidification which could exceed 0.5 pH units. In contrast, [Ca2+]e remained unaffected by anoxia. Addition of 2-5 mM cyanide (CN-) to oxygenated Hepes-buffered saline evoked an increase in PO2 in the VRG from 100 to more than 300 mmHg. The effects of CN- on respiratory activity, aKe and pHe were almost identical to those during anoxia. In oxygenated, CO2/bicarbonate-free solutions of different pH, however, an increase in pHe in the VRG led to a decrease in respiratory frequency, whereas a fall of pHe produced a frequency acceleration. A rise of aKe in the VRG by more than 2 mM as induced by superfusion of a 7 mM K+ solution led to a sustained increase of respiratory frequency. The results indicate that blockade of aerobic metabolism does not severely perturb K+ and Ca2+ homeostasis and that the biphasic response to anoxia is not directly related to the observed changes in PO2, aKe, pHe, or [Ca2+]e. In the respiratory network of neonatal mammals, CO2 might provide a stimulus for long-term maintenance of respiratory activity under oxygen depletion.