Background: The Bay of Marseilles is under the influence of a large urban concentration and its maritime activities. All of them discharge compounds (hydrocarbons, excess nutrients, heavy metals, chemicals, etc.) that can alter the marine ecosystem. To investigate whether ultraphyto- plankton ( , 10 m m) could be used as biosensors for their own ecosystem, a 2-year survey was conducted in the Bay of Marseilles. Methods: Seven stations monitored water mass and po- tential anthropic effects in the bay. Seawater samples were collected monthly or bimonthly at three depths, prefiltered, fixed, and kept in liquid nitrogen until flow cytometric analysis. Results: Five categories were created: Prochlorococcus, picoeukaryotes ( , 2 m m), nanoeukaryotes I (2- 6 m m), nanoeukaryotes II (6 -10 m m), and Synechococcus ( , 1.5 m m). Artificial neural network analysis (Kohonen self-or- ganizing maps) produced the same number of clusters as cluster analysis with Winlist software (Verity Software House). Conclusions: In addition to the wide variabilities in abun- dance and biomass, there were a strong seasonal signal and sporadic events. Lessons are derived from this study for future monitoring of marine microorganisms. Cytometry 44:247-256, 2001. © 2001 Wiley-Liss, Inc.
