In the Bay of Brest (France), benthic fluxes of silicic acid (DSi) are thought to be strongly involved in diatom dominance and their intensity depends on the density of an invasive gastropod, Crepidula fornicata. While the proliferation of this gastropod is likely to support diatom dominance, local fishermen have proposed its removal as it perturbs the development of the native Great Scallop, Pecten maximus. However, such a removal may be associated with potentially harmful algal blooms. The response of the phytoplankton community to modifications of the benthic fluxes of DSi, resulting from a further proliferation or the removal of C fornicata, was examined in a mesocosm nutrient enrichment experiment over a period of 16 days in the spring of 2004. Six mesocosms (5.4 m(3)) were filled using a natural planktonic assemblage from the Bay and received N, P and Si additions at the stoichiometric ratio of the in situ benthic fluxes (Si:N:P=16:16:1), with a gradient of enrichment ranging from 0 to almost 4 times the maximal benthic flux measured in the Bay. The diatom Chaetoceros (mainly C. debilis) rapidly outgrew non-siliceous species of microphytoplankton in all enclosures although kinetic experiments provided direct evidence of Si limitation. This dominance of diatoms at DSi concentrations <2 mu M highlighted that the residual DSi concentration measured daily in the mesocosms is not a reliable criterion to predict diatom dominance in marine ecosystems whereas the dependence of the Si uptake rate on fertilization rates suggests that emphasis should be placed on daily nutrient addition. Quantitative modifications of nutrient additions only changed the magnitude of the diatom bloom and high fertilization rates may lead to high sedimentation fluxes of organic material, which can strongly harm the entire benthic community. Finally the diatom dominance at low fertilization rates raised questions about the role of benthic fluxes in preventing harmful algal blooms within the Bay system.