Accidental microbiological contamination of distributed water (referred to as "contamination") was defined as the presence of at least 20 colony-forming fecal bacteria (Escherichia coli and enterococci) in analyses conducted on samples (100 mL test samples) collected on the same day from the same distribution unit (DU). 2,739 "contaminations" occurred in small DUs (<2,000 users served) between 2003 and 2004 in the 90 departments of mainland France, excluding Paris and its inner suburbs. After standardizing for the distribution of DUs by size (divided into 3 classes), we explored the correlations between the number of "contaminations" per department (N_Pols) and meteorological, hydrological, topographical, and geological indicators, as well as indicators of water sanitation management.N_Pols is linked to four covariates: (i) precipitation, very significantly, (ii) the extent of karst areas, (iii) the prevalence of disinfection, and (iv) that of measures to protect water catchments. An excess of 100 mm of annual precipitation corresponds to a 30% increase in the risk of "pollution." A 10% increase in the proportions of karstified areas, resources without protection zones, and water supply units lacking disinfection is accompanied by a 10% increase in the frequency of "pollution" events. The correlations are consistent across the three water supply unit size classes and align with expert expectations. It would be worthwhile to repeat these studies over time to strengthen the interpretation regarding causality and, where appropriate, to track changes in risk and its determinants. (R.A.)