Dispersant application is an oil spill response technique. To evaluate the environmental cost of this operation in nearshore habitats, the experimental approach conducted in this study exposed juvenile golden grey mullets (Liza aurata) for 48 hours to chemically dispersed oil (simulating, in vivo, dispersant application), to dispersant alone in sea water (as an internal control of chemically dispersed oil), to mechanically dispersed oil (simulating, in vivo, natural dispersion), to the water-soluble fraction of oil (simulating, in vivo, an oil slick confinement response technique) and to sea water alone (control condition). Biomarkers such as fluorescence of biliary polycyclic aromatic hydrocarbon (PAH) metabolites, total glutathione liver content, EROD (7-ethoxy-resorufin-O-deethylase) activity, liver antioxidant enzyme activity, liver lipid peroxidation and an innate immune parameter (haemolytic activity of the alternative complement pathway) were measured to assess the toxicity of dispersant application. Significant responses of PAH metabolites and total glutathione liver content to chemically dispersed oil were found, when compared to water-soluble fraction of oil. As it was suggested in other studies, these results highlight that priority must be given to oil slick confinement instead of dispersant application. However, since the same patterns of biomarkers responses were observed for both chemically and mechanically dispersed oil, the results also suggest that dispersant application is no more toxic than the natural dispersion occurring in nearshore areas (e.g. waves). The results of this study must, nevertheless, be interpreted cautiously since other components of nearshore habitats must be considered to establish a framework for dispersant use in nearshore areas.