Mutagenicity of nine epoxy resin hardeners was examined by a fluctuation test modified by Gatehouse. The test was performed by using Salmonella typhimurium TA98 with a metabolic activation system. In our laboratory, the results of the fluctuation test were compared with the results obtained by the previously mentioned Ames preincubation method. Six out of nine epoxy resin hardeners showed mutagenic activity in both the fluctuation test and Ames preincubation method, but one out of the nine was negative in both test systems. Two out of the nine were positive by either of the two testing systems. The fluctuation test is disadvantageous in that it is marginally slower and requires slightly more labor than the Ames test and furthermore it is difficult to increase the amount of microsome because of background interference. These disadvantages, however, are somewhat offset by the advantages that small organs such as urinary bladder can be used instead of liver cells and that a small amount of microsome can be employed for metabolic activation. This test is also suitable for testing aqueous samples containing low levels of mutagen.