A previously described neutral red uptake-inhibition assay on cultured fathead minnow (FHM) fish cells revealed a good correlation with fish toxicity data. The assay was made more sensitive by reducing the number of cells, by using a longer treatment period, and by simultaneous treatment of the cells with sodium dodecyl sulphate and L-buthionine (S,R)-sulphoximine. The fluorimetrically quantified protein content was then used as the endpoint. The results were compared with toxicity data obtained by the Microtox® method with the bacterium, Vibrio fischeri. A series of 82 surface-water samples were investigated. Cell quantity-dependent and sample concentration-dependent reductions in total protein content were observed. In all, 32 samples were toxic in the FHM assay. Of the seven samples that were shown to be toxic in the Microtox assay, only three were also toxic in FHM cells. No linear relationship was found for the toxicity results obtained with the two methods. Further research to explain the significance of the many positive responses and to identify possible confounding factors will increase the reliability of the sensitised FHM cell assay.