Acrylamide monomer (AA), injected into male mice at the maximum tolerated dose of 5 x 50 mg/kg (24-h intervals), significantly increased the specific-locus mutation rate in certain poststem-cell stages of spermatogenesis, but not in spermatogonial stem cells. Germ-cell stages in which the treatment induced dominant lethals - namely, exposed spermatozoa and late spermatids (number of surviving offspring only 3% and 27%, respectively, of those in concurrent controls) - jointly yielded the highest frequency of specific-locus mutations. AA thus conforms to Pattern 1 in our earlier classification of chemicals according to the spermatogenic stage at which they elicit maximum response (Russell et al., 1990). No specific-locus mutations were observed among 17,112 offspring derived from exposed spermatogonial stem cells, a result which rules out (at the 5% significane level) an induced mutation rate greater than 2.3 times the historical control rate. A sustained high productivity in matings made for several months following week 3 indicates that there is no significant spermatogonial killing and that cell selection is presumably not the explanation for the negative result. On the basis of genetic and/or cytogenetic evidence, the mutations induced postmeiotically by AA were 'large lesions' (multi-locus), while one of 2 recovered from exposure of differentiating spermatogonia is probably a small lesion. An earlier survery of mammalian mutagenesis results led us to conclude that, regardless of the classification of a chemical according to the stage at which it elicits its maximum response, the nature of mutations is determined by the germ-cell stage in which they are induced (Russell et al., 1990). The AA results on lesion size and on distribution of mutations among the loci fit the general pattern.