Closed cultures of marine phytoplankton were established under variable conditions of CO2 concentration, temperature, growth rate (by light limitation), and pH (but with nearly identical [CO2aq]) in order to assess the relative influence of these variables on the extent of carbon isotope fractionation relative to dissolved inorganic carbon sources. Culture biomass was not allowed to increase beyond levels that would significantly affect the dissolved carbon system in the closed cultures. In experiments with Skeletonema costatum and Emiliania huxleyi, increasing CO2 concentrations led to increased carbon isotope discrimination (resulting in organic matter progressively depleted for &dgr;13C, i.e., a greater, more negative &egr;p). &egr;p values for E. huxleyi were 8--10%0 less than for S. costatum under identical conditions. For the S. costatum cultures, there was nearly a 20%0 range in [CO2aq]-dependent &egr;p. The effect was nonlinear with a leveling off at high[CO2aq]. Over a pH range of 7.5--8.3 but at a constant [CO2aq] there was a variation in carbon isotope fractionation by S. costatum of about 9%0 with a minimum at pH 7.8--7.9. There was a temperature effect of ~8%0 on fractionation even after equilibrium temperature dependency of &dgr;13C of CO2aq was taken into account. No growth rate effect was found for S. costatum over a modest range of growth rates. Culture experiments used to determine the carbon isotope fractionation by phytoplankton species must be conducted under well-defined conditions of temperature, pH, and CO2 concentrations. Hindcasts of ancient atmospheric pCO2 from measurements of &dgr;13C of organic carbon in marine sediments will require careful calibration because of the variety of possible factors that influence &dgr;13Corg.