Microcosms, each consisting of 2 L natural surface seawater maintained in 2.3-L glass bottles, were immersed at a depth of 6 metres. The renewal of ten percent of microcosm volumes was carried out every other day. Phytoplankton-containing seawater was used for renewal (previously filtered through 25-, 50- or 200-μm cut-off). Phytoplankton community pigment analysis (by HPLC) and flow cytometry analysis were performed. After 13 days, data exhibited phytoplankton characteristics in microcosms in the same range as that of the natural surrounding sea water over the same period. Furthermore, in these microcosms, a negative correlation was observed between the filtration cut-off used for renewal water, and the eukaryote cell count. Herbicides were tested as commercial mixtures at 1, 10 and 100 μg.L-1 active principal. Both Frontier® (dimethenamid) and Basamais® (bentazon) induced significant modifications of the phytoplankton populations at every concentration tested. Such results suggest a possible disturbance in polluted coastal areas.