Cotton plants were grown in CO2-controlled growth chambers in atmospheres of either 35 or 65 Pa CO2. A widely accepted model of C3 leaf photosynthesis was parameterized for leaves from both CO2 treatments using non-linear least squares regression techniques, but in order to achieve reasonable fits, it was necessary to include a phosphate limitation resulting from inadequate triose phosphate utilization. Despite the accumulation of large amounts of starch (>50 g m−2) in the high CO2 plants, the photosynthetic characteristics of leaves in both treatments were similar, although the maximum rate of Rubisco activity (Vcmax), estimated from A versus Ci response curves measured at 29°C, was ∼10% lower in leaves from plants grown in high CO2. The relationship between key model parameters and total leaf N was linear, the only difference between CO2 treatments being a slight reduction in the slope of the line relating Vcmax to leaf N in plants grown at high CO2. Stomatal conductance of leaves of plants grown and measured at 65 Pa CO2 was approximately 32% lower than that of plants grown and measured at 35 Pa. Because photosynthetic capacity of leaves grown in high CO2 was only slightly less than that of leaves grown in 35 Pa CO2, net photosynthesis measured at the growth CO2, light and temperature conditions was approximately 25% greater in leaves of plants grown in high CO2, despite the reduction in leaf conductance. Greater assimilation rate was one factor allowing plants grown in high CO2 to incorporate 30% more biomass during the first 36 d of growth.