We tested the applicability of the carbon isotopic composition of C37:2 alkenones (delta13C37:2) as a proxy for dissolved carbon dioxide CO2(aq) in oceanic surface waters. For this purpose we determined delta13C37:2 in suspended particulate organic matter (POM) and surface sediments from the South Atlantic. In opposite of what would be expected from a diffusive CO2 uptake model for marine algae we observed a positive correlation between 1/[CO2(aq)] and the isotopic fractionation (epsilon p) calculated from delta13C37:2. This clearly demonstrates that CO2(aq) is not the primary factor controlling epsilon p at the sites studied. On the other hand we found a negative correlation between epsilon p and the phosphate concentration in the surface waters (0-10 m) supporting the assumption of Bidigare et al. (1997) that epsilon p is primarily related to nutrient-limited algal growth rather than to [CO2(aq)]. Reconstructing past CO2(aq) levels from delta13C37:2 thus requires additional proxy information in order to correct for the influence of haptophyte growth on the isotopic fractionation. In the eastern Angola Basin, we previously used delta15N of bulk organic matter as proxy for nutrient-limited growth rates. As an alternative the Sr/Ca ratio of coccoliths has been recently suggested as growth-rate proxy which should be tested in future studies.