||Two common tallgrass prairie species, Andropogon gerardii, the dominant C-4 grass in this North American grassland, and Salvia pitcheri, a C-3 forb, were exposed to ambient and elevated (twice ambient) CO2 within open-top chambers throughout the 1993 growing season. After full canopy development, stomatal density on abaxial and adaxial surfaces, guard cell length and specific leaf mass (SLM; mg cm(-2) were determined for plants in the chambers as well as in adjacent unchambered plots. Record high rainfall amounts during the 1993 growing season minimized water stress in these plants (leaf xylem pressure potential was usually > -1.5 MPa in A. gerardii) and also minimized differences in water status among treatments. In A. gerardii, stomatal density was significantly higher (190 +/- 7 mm(-2); mean +/- s.e.) in plants grown outside of the chambers compared to plants that developed inside the ambient CO2 chambers (161 +/- 5 mm(-2)). Thus, there was a significant ’chamber effect’ on stomatal density. At elevated levels of CO2, stomatal density was even lower (P < 0.05; 121 +/- 5 mm(-2)). Most stomata were on abaxial leaf surfaces in this grass, but the ratio of adaxial to abaxial stomatal density was greater at elevated levels of CO2. In S. pitcheri, stomatal density was also significantly lower when plants were grown in the open-top chambers (235 +/- 10 mm(-2) outside vs. 140 +/- 6 mm(-2) in the ambient CO2 chamber). However, stomatal density was greater at elevated CO2 (218 +/- 12 mm(-2)) compared to plants from the ambient CO2 chamber. The ratio of stomata on adaxial vs. abaxial surfaces did not vary significantly in this herb. Guard cell lengths were not significantly affected by growth in the chambers or by elevated CO2 for either species. Growth within the chambers resulted in lower SLM in S. pitcheri, but CO2 concentration had no effect. In A. gerardii, SLM was lower at elevated CO2. These results indicate that stomatal and leaf responses to elevated CO2 are species specific, and reinforce the need to assess chamber effects along with treatment effects (CO2) when using open-top chambers.