A theoretical model of CO2aq-dependent phytoplankton carbon isotope fractionation (&egr;p) and abundance (&dgr;13Corg) is compared to observed isotopic trends with temperature and [CO2aq] in the ocean. It is shown that the model's &dgr; 13Corg response to surface ocean temperature and to [CO2aq] can simulate observed trends when the other independent model variables (phytoplankton cell growth rate, cell size, cell membrane CO2 permeability, and enzymatic isotope fractionation) are held at realistic, constant values. The possible contribution made by each of these variables to the residual scatter in &dgr;13Corg about its trends with temperature and [CO2aq] is quantified, thus estimating a maximum isotopic sensitivity to changes in each of these variables. The model response to growth rate and especially cell size, however, appears to be unrealistically high. This may occur because the net isotopic effects of such factors may be attenuated through dependent and isotopically offsetting variations among variables. The model's indicated sensitivity to such factors as CO2 permeability, enzymatic fractionation, cell size, and cell surface area/volume provides mechanisms whereby changes in species composition can play a significant role in affecting observed variations in oceanic &dgr;13Corg. Overall, the model is consistent with earlier suggestions that marine &dgr;13Corg and &egr;p variability can be explained by carbon isotope fractionation evoked by CO2aq-dependent phytoplankton. This has important implications for interpreting carbon isotopic variability encountered in plankton and their organic constituents in the present-day ocean and in the marine sedimentary record.¿ 1997 American Geophysical Union