Three differing bacterial toxicity assays were compared: the 'Microtox' test, (Photobacterium phosphoreum luminescence inhibition assay), the 'oxygen consumption of activated sludge' assay (ISO 8192), and the 'Glucose U-14 C mineralization' assay (the rate of release of 4 CO2 by 'Escherichia col£). Metals, amines, halogenated alcans, chlorophenols, aromatic hydrocarbons, surfactants, and pesticides were screened for their toxic activity.Results showed satisfactory repeatability of the three bacterial assays with variation coefficients between 5 and 32%. The 'Microtox' assay was the most sensitive test evaluated under our conditions. The lower sensitivity of the 'oxygen consumption' assay may have been due to high concentrations of substrates which modify toxicant bioavailability, and also to a high biomass/toxic substances ratio. The 'Glucose U-4 C mineralization' assay was selective, and low in sensitivity; but the specific species used in this test -Escherichia coli -may have been responsible for this selectivity.The 'Microtox' test appears to be well adapted to the detection of aquatic environmental pollution, and to the toxicity screening of complex solid waste effluents and/or leachates. The 'oxygen consumption' assay can be advantageously used to measure the impact of sewage on activated sludge in biological treatment plants. The 'Glucose U-' 4 C mineralization' assay, which does not require high biomass, can be useful for in situ studies using field microorganisms.