We describe here historical negative control data bases for gene locus mutation assays at the thymidine kinase locus and hypoxanthine-guanine phosphoribosyl transferase locus in TK6 and AHH-1 human lymphoblasts, respectively. Protocols have been designed which minimize the variability among independent experiments and thus facilitate the use of historical negative control data bases for assay analysis. The historical negative control data bases for both cell lines can be accurately modeled as gamma or Poisson distributions; confidence limits can be calculated from these distributions. We describe and justify a mutagenicity assay analysis procedure which uses a comparison to the concurrent negative control cultures via a t-test in conjunction with a comparison to the historical negative control data base. The incorporation of a comparison to the historical negative control data base allows the use of a higher confidence level without substantially sacrificing the sensitivity of the mutation assays. The analysis of the mutagenicity of saccharin in TK6 cells is presented as an example; saccharin was found to be nonmutagenic under the conditions tested.