The carbon isotopic composition (813C) of the marine diatom PhaeodactylurY1 tricomutum was measured over a series of growth rates (p) in a chemostat system in which both the Si3C and the concentration of aqueous CO, [CO,(aq)] were measured. CO,(aq) ranged from 0.64 to 35 pmol kg-l and growth rates from 0.5 to 1.4 d-l. eP, the biological fractionation factor associated with carbon fixation, was found to bc a nonlinear function of @O,(aq), contrary to the predictions of a model that assumes that CO, enters the cell by passive diffusion. The experimental results suggest that active uptake of bicarbonate does not account for the nonlinearity of the relationship and that inorganic carbon enters the cell as CO,. The data are very well described by a theoretical model that assumes that P. tricornutum regulates the CO, concentration in its cytoplasm so as to minimize the energy required to concentrate CO, at the site of carboxylation. This is probably achieved by active uptake of CO, or by conversion of bicarbonate to CO, by an external carbonic anhydrase followed by transport of the CO, into the cell via either active transport or passive diffusion. Based on the model and data, @ZO,(aq) = 0.225 X [(26.8 -a)/(~, -5.5)] kg d-' pmol-l. This equation accounts for 92% of the variance in the @ZO,(aq) data. The model has potential utility for estimating phytoplankton growth rates in field studies without incubations and has important implications for the estimation of ancient CO,(aq) from the 613C of preserved organic compounds.There is general agreement that the 13C : 12C ratio of the organic carbon in plants reflects both the physiological condition of the plants and certain aspects of the environment at the time the organic matter was formed. Farquhar et al. (1982), for example, demonstrated that the combined effects of diffusion of CO, into a plant and isotopic discrimination by the primary carboxylating enzyme ribulose bisphosphate carboxylase oxygenase (Rubisco) would lead to approximately a linear relationship between the 613C of plant organic matter and the ration of the internal-to-external CO, concentrations. Fry and Wainright (1991) first recognized that both aqueous CO, concentrations and microalgal growth rates would influence the 613C of phytoplankton, and Rau et al. (1992) pointed out that negative correlations between the S"C of suspended particulate organic matter and concentrations of aqueous CO, could be explained by variations in phytoplankton demand for CO,. Francois et al. (1993) showed that if passive diffusion dominated carbon transport into phytoplankton cells, there should exist a linear relationship between the ST of phytoplankton organic carbon and the ratio of the carboxylation rate to external aqueous CO, concentration. Building on this earlier work, Goericke et al. (I 994) and Laws et al. (1995) derived equations that predicted that the S'"C of microalgal organic carbon should be