Data from experiments on the induction of specific-locus mutations in model systems are utilized in genetic risk assessment to estimate potential adverse effects in the human population. In such assessments with radiation or chemical mutagens, the following information is required: (1) spontaneous and induced forward-mutation frequencies, (2) dose-response curves for the overall induction of specific-locus mutations, (3) genetic characterization of spontaneous and induced mutations, and (4) dose-response curves for the different genotypic classes. Specific-locus assays in most eukaryote assay systems provide only portions of the information required for genetic risk assessment. In recognition of the need for a more comprehensive data base, a model system was developed for specific-locus studies in Neurospora crassa. The adenine-3 (ad-3) specific-locus assay was modeled after the 2 gene, morphological specific-locus assay in the dilute-short-ear region of the mouse, and it detects forward-mutations at two closely linked loci: ad-3A and ad-3B. The ad-3 assay system has provided precise dose-response curves not only for inactivation, but also the overall induction of ad-3 mutations. The utilization of this assay in experiments with radiation or chemical mutagens has provided a data base on the induction and genetic characterization of specific-locus mutations that is unique among eukaryotic organisms. In this assay, gene/point mutations, multilocus deletion mutations, and 3 different classes of multiple-locus mutations can be identified. The latter consist of specific-locus mutations associated with recessive lethal mutations located either closely linked to the ad-3 region or elsewhere in the genome. The overall data base on the heterozygous effects of X-ray-induced ad-3 mutations demonstrates that such effects are allele specific, genotype specific, and locus specific. There are probably a variety of mechanisms by which the heterozygous effects of individual allelic mutations at different genetic loci can be affected. In conclusion, unless the frequencies of all of the different classes of induced specific-locus mutations are determined, and utilized in genetic risk assessment exercises, the risk of human exposure to environmental mutagens may be grossly underestimated.