Abstract The effects of N-ethyl- N′-nitro- N-nitrosoguanidine (MNNG) on the induction of micronuclei were examined in mouse fetuses exposed in utero. By this study, MNNG was proved to be mutagenic in vivo. The frequency of micronucleated erythrocytes (MNEs) in fetal liver peaked at 18 h after a single intraperitonela injection into pregnant mice on day 13 of gestation. Then, to examine the effects of administration routes on the induction of micronuclei, the chemical was given by various routes, and the percentage of MNEs (%MNEs) in fetuses were examined 18 h after treatment. The %MNEs after administration of MNNG intraperitoneally, subcutaneously, intravenously, and orally was 4.7, 1.9, 0.8 and 0.3, respectively. The control value was 0.3. In the intraperitoneally treated mice, %MNEs for fetuses in the uterine horn olcated nearer the injection site was higher than that in the other. In addition, in the intraperitoneally treated mice, there was a tendency for the higher %MNEs to occur in the fetuses located near the injection site. Together with the results on the distribution of MNNG in mice (Frei and Lawley, 1976), these findings suggest that MNNG might be inactivated in the maternal systemic circulation and that the agent which induces micronuclei might be distributed to the fetuses by diffusion.