During field measurements in a bloom of Emiliania huxleyi in the North Sea in 1993, an apparently inconsistent combination of observations was measured: (1) fCO2 was lower in the center of the bloom than in the surrounding nonbloom areas and undersaturated with respect to the atmosphere in both cases, (2) within the bloom, enhanced sedimentation of coccoliths-containing fecal pellets was observed, (3) a large atmospheric sink of 1.3 mol C m-2 was derived, and (4) in the same bloom a positive correlation between CaCO3 and fCO2 was observed, which was interpreted as an increase of fCO2 during production of CaCO3. In order to resolve the inconsistency between observations (1, 2, 3) and 4 a one-dimensional three-layer model was constructed. A positive correlation between CaCO3 and fCO2 was obtained when the model was parameterized with data obtained from field and mesocosm studies. The correlation is a feature of the decay phase of a bloom and represents a decrease of fCO2 with a decrease of CaCO3. Thus it represents the dissolution and sedimentation of CaCO3 rather than its production. Having resolved the ambiguity within the field data by adding the dimension of time in the model, blooms of E. huxleyi can be identified as sinks for atmospheric carbon dioxide. This sink is a function of the calcification to photosynthesis (C:P) ratio of the nitrate-using phytoplankton and is maximal when the C:P ratio is 0.42 (that is, E. huxleyi constitutes 97% of the nitrate-using phytoplankton). Rather than using the model for making accurate predictions about the magnitude of this sink, a sensitivity analysis was performed to give a range of magnitudes for the range of parameter values that were obtained during previous studies. Furthermore, gaps were identified in the current knowledge of carbon fluxes within blooms of E. huxleyi. ¿ 2001 American Geophysical Union