Abstract Cytological and fertility tests were performed in F 1 male mice derived from different germ-cell stages of male parents treated with cyclophosphamide (350 mg/kg body weight). The objectives of the present experiment were: (1) to determine the sensitivity of the male germ-cell stages to the induction of translocations by the compound, and (2) to characterize translocation configurations in F 1 and F 2 males, in order to obtain information about the pattern of chromosome breakage induced and its transmission to subsequent generations. Of 508 F 1 males studied, 39 were partially sterile and 9 were fully sterile. The group of males conceived 8–21 days after treatment contained by far the highest proportion of partially sterile animals (30%). It was also the only group in which totally sterile animals (11%) were found. Of 25 semisterile males from this group, 24 gave evidence of translocations when spermatocytes were scored at diakinesis. The translocation frequencies in F 1 derived from treated spermatozoa and spermatocytes were 14 and 1%, respectively. No translocations were detected cytologically 6 semisterile males derived from treated spermatogonial stages. These results indicate that spermatid stages are especially sensitive to the mutagenic action of cyclophosphamide. In 21 of the 31 semisterile translocation males (68%), the majority of the spermatocytes contained 18 bivalents plus a ring-of-four configuration, indicating that both breakpoints were relatively centrally located; and in several of these males, the frequency of cells with rings was close to 100%. In another 9 F 1 males (29%) the predominant multivalent configuration was a chain-of-four, indicating one of the breakpoints to be relatively more terminally located; and in one male (3%), the majority of cells had two unequal bivalents, indicating both breakpoints to be fairly close to the ends of the chromosomes involved. Determination of centromere positions by the use of C-banding showed that chain-of-four configurations in any one male were predominantly of a given type. Associations of more than one translocation were observed in 11 of 31 translocation males, but in a very low frequency (about 1% of the spermatocytes). The results indicate that cyclophosphamide induces a relatively random pattern of chromosome break location. In this respect this mutagen more closely resembles X-rays than it does certain chemicals for which there isf evidence that breakpoint position is frequently subterminal. Our data indicate that the pattern of configurations observed in F 1 was in general transmitted to F 2.