Undisturbed tallgrass prairie, dominated by the C-4 grass Andropogon gerardii, was exposed to ambient and elevated (double ambient) levels of atmospheric CO2 in large open-top chambers throughout the 1991 and 1992 growing seasons. Responses in leaf xylem pressure potential (psi), net photosynthesis (A), and stomatal conductance (g) were measured in both years for A. gerardii grown within chambers and from adjacent field plots. In 1992, maximum photosynthetic capacity (A(max)), apparent quantum requirement (Q(r)), the photosynthetic light compensation point (LCP), and dark respiration (R(d)) were also measured. Midday psi was significantly higher in plants grown at elevated CO2 in both years, and seasonally averaged psi was 0.48-0.70 MPa lower in 1991 (a dry year) than 1992 (a wet year). In 1991, A and g were significantly higher (regardless of measurement CO2 level) in plants grown at elevated vs. ambient CO2. These increases were measured in well-watered plants insuring that these plants differed only in CO2 growth conditions and previous exposure to low psi. Increased A at elevated CO2 occurred (as much as 7.1 mu mol m(-2) s(-1)) over a broad range of temperatures (17-35 C), but the temperature optimum for A was similar at both 350 and 700 mu L L(-1) CO2. In 1992, no differences in A, A(max), Q(r), LCP, or R(d) were detected when ambient and elevated CO2 plants were compared. In plants collected from field plots, R(d), LCP, and leaf N were significantly higher than in plants within the chambers indicating that a chamber effect exists for these parameters. In both years, g was significantly reduced (21%-51%) when measured at 700 vs. 350 mu L L(-1) CO2. Peak aboveground biomass was increased at elevated CO2 in 1991 but not in 1992. These data indicate that for C-4 grasses, effects of elevated CO2 may only be detectable in years with significant water stress, a common occurrence in the central North American tallgrass prairies.