An organism's critical oxygen tension (Pcrit) reflects its ability to extract environmental O2. Consequently, Pcrit has been used as an indicator of hypoxia tolerance in aquatic animals. The relationship between Pcrit and hypoxia tolerance and hypoxic metabolic responses, however, remains incompletely understood. Among several species of sculpin fishes (superfamily Cottoidea), we previously demonstrated a correlation between Pcrit and hypoxia tolerance, as measured as the time required for 50% of a group of fish to lose equilibrium (LOE50) at a water PO2 of 6.4Torr. In the present study, we further investigated the relationship between Pcrit, hypoxia tolerance, and hypoxic metabolic responses by examining the effects of hypoxia exposure at a fixed percentage of Pcrit (30%; termed relative hypoxia exposure) on LOE50 and metabolic responses in brain, liver, and white muscle in three sculpin species that differ in Pcrit and hypoxia tolerance at 6.4Torr. We also assessed the tissue metabolic responses underlying hypoxic loss of equilibrium (LOE). The species, from most to least hypoxia-tolerant at 6.4Torr and from lowest to highest Pcrit values, were the tidepool sculpin (Oligocottus maculosus), staghorn sculpin (Leptocottus armatus), and silverspotted sculpin (Blepsias cirrhosus). If Pcrit predicts hypoxia tolerance, then we expected similar LOE50 values and similar tissue metabolic responses across all species during relative hypoxia exposure. LOE50 values were similar in staghorn sculpins and tidepool sculpins, but not in silverspotted sculpins, which had a comparatively lower relative hypoxia LOE50 value. Thus, Pcrit, and consequently the ability to extract environmental O2, cannot predict hypoxia tolerance in all species, at least at a water PO2 of 30% of Pcrit. During relative hypoxia exposure, tissue lactate accumulation and ATP levels were similar between species, suggesting that the ability to extract environmental O2 is an important determinant of cellular energy status and reliance on anaerobic glycolysis in hypoxic sculpins. However, whereas tissue glycogen content and utilization were similar between tidepool sculpins and staghorn sculpins, there were lower normoxic levels and greater hypoxic depletion in silverspotted sculpins, potentially explaining their poorer relative hypoxia tolerance. In all species, LOE was associated with depletion of brain [ATP]. Overall, caution is warranted when Pcrit is used as an indicator of hypoxia tolerance, especially when considering temporal aspects of hypoxia tolerance and related metabolic characteristics (e.g. glycogen availability). Ideally, comparative studies of hypoxia tolerance should feature multiple measures (e.g. Pcrit and LOE50) in order to assess the overall responses of fishes to hypoxia.