Specific-locus experiments have previously shown melphalan to be mutagenic in all male germ-cell stages tested and particularly so in early spermatids. All but 2 of 24 specific-locus mutations recovered were tested genetically, cytogenetically, and/or molecularly. At least 12 of 15 tested mutations recovered from postspermatogonial stages but only 1 of 7 mutations recovered from stem-cell or differentiating spermatogonia gave evidence of being deletions or other rearrangements. Melphalan-induced mutations, thus, confirm the pattern of dependence of mutation structure on germ-cell stage that had been shown earlier for other chemicals. Results of the present investigation illustrate the capabilities of combined genetic, cytogenetic, and molecular analyses for characterizing the nature of specific-locus mutations. Fine-structure molecular mapping of long regions surrounding specific loci has been greatly facilitated by the availability of genetic reagents (particularly, deletion complexes) generated in specific-locus experiments over the course of decades. Reciprocally, this mapping permits increasingly detailed characterization of the nature of lesions induced by mutagenic exposures of germ cells, adding great powers for qualitative analysis of mutations to the specific-locus test. Cytogenetic and genetic investigations also provide evidence on lesion type, especially for loci at which mutations cannot yet be analyzed molecularly. Melphalan, like chlorambucil, can generate many mutations, a high proportion of which are deletions and other rearrangements, making this chemical valuable for generating mutations (at any locus) amenable to molecular access.