The effects of variations in volume of solution of toxicant, aggregation, temperature, and change in condition of test animals during time of holding in the laboratory were investigated with regard to possible influences on bio-assay results. The western gambusia, Gambusia affinis affinis (Baird and Girard) was used as the test animal. In spite of relatively favorable experimental conditions, it seems that estimates of relative toxicity are affected by uncontrolled variations which greatly limit dependability.