A new technology for treating waters contaminated with acid mine drainage involves the dissolution of limestone particles using carbon dioxide at pressures above ambient. Because of the fish health risks associated with episodes of high carbon dioxide levels in treated waters, we subjected three species of fish, brook trout (Salvelinus fontinalis), slimy sculpin (Cottus cognatus), and blacknose dace (Rhinichthys atratulus), to 24 h exposures of elevated dissolved carbon dioxide (CO₂) at three levels, ranging from 1.0 (low) to 6.3 (high)%, under laboratory conditions. We measured blood physiological variables as well as behavior, including feeding responses, before, during, and after exposure. Physiological responses differed by species, but all species had elevated hematocrits after 1 h of exposure. Brook trout hematocrits were higher at medium and high levels of CO₂ than in a control group (0.0% CO₂) after 24 h of exposure. Slimy sculpin hematocrits were higher in medium- and high-level exposure groups than in controls after 1 h, but not after 24 h, of exposure. Blacknose dace hematocrits were higher in all three exposure groups than in controls after 1 h but only in medium-level exposure groups after 24 h. Brook trout plasma glucose was significantly higher in medium- and high-level exposure groups after 1 h, and in the high-level group after 24 h, than in controls. Slimy sculpin plasma glucose was not significantly different in elevated CO₂ exposure groups from that of controls throughout exposure. Branchial ventilation was significantly greater in all species at elevated CO₂ during exposure, indicating stress; however, no difference was observed between treatment and control groups of blacknose dace after 24 h, indicating acclimation. Pectoral fin beats and cough rates were not consistently related to CO₂ exposure throughout the study. Brook trout had the longest lasting reaction to stress at lower levels of CO₂ among the three species tested. Many of the 11 observed behavioral variables, related to swimming, feeding, social, and illness factors, were affected by elevations of dissolved CO₂. Two to seven behavioral variables (18–64% of those measured) were affected by treatment level of dissolved CO₂ with a trend by species for the number of variables affected: brook trout>blacknose dace>slimy sculpin. However, behavioral sensitivity to treatment level was greatest in blacknose dace. Recovery to pre-treatment activity rates for most behavior patterns (including feeding) was observed 24 h after cessation of exposure in all three species. Recovery was independent of treatment level, was most rapid in blacknose dace, and was slowest in brook trout. Overall, slimy sculpin was least affected behaviorally by elevated CO₂. Although all three species showed stress response and changes in behavior at moderate levels of CO₂ (≥2%), brook trout and blacknose dace showed evidence of ability to avoid harmful CO₂ levels by swimming out of affected waters, whereas the slimy sculpin showed minimal behavioral changes despite remaining in place during exposure. Thus, predation risk and other sources of mortality seem minimal in the event of technological malfunction at a stream treatment site involving the use of CO₂ under pressure.