Quinolones are one of the most important group of synthetic antibiotics used in aquaculture. We studied the single substance and mixture toxicity of ten quinolones using a long term bioluminescence inhibition assay with the marine bacterium Vibrio fischeri as the test organism. All tested quinolones are highly toxic to the test organism with EC50 values ranging from 14 µg/l for ofloxacin to 1020 µg/l for pipemidic acid. Adapting the approach outlined in EEC directive 93/21/EEC to these results, all but one of the ten quinolones belong to the group classified as being 'very toxic to aquatic organisms' (EC50 below 1 mg/l). On the basis of the concentration-response relationships of the single compounds, the mixture toxicity of the ten compounds was estimated by the concepts of concentration addition and independent action. Complete concentration-response relationships were experimentally recorded for the quinolone mixture in three different mixture ratios, based on the relative toxicity of the components (EC50, EC1 and NOEC). The results show that the mixture toxicity of the quinolones is best predictable by concentration addition, whereas independent action underestimates the toxicity of the mixture. As the quinolones have an identical specific mechanism of action (the inhibition of bacterial gyrases), these results are in agreement with the pharmacological assumptions that form the basis of the concept of concentration addition. It is therefore concluded, that concentration addition can be useful for hazard assessment procedures of mixtures of similarly acting compounds. One important implication of this concept is that even mixture components that are present only at their individual no observed effect concentrations (NOECs) contribute to the overall toxicity of the mixture. Under these conditions more than 99% effect of the quinolone mixture are observed. This result emphasises the unsuitability of NOECs as an approximation of a 'safe' concentration.