Parental exposure of mice to radiation and chemicals causes a variety of adverse effects (e.g., tumors, congenital malformations and embryonic deaths) in the progeny and the tumor-susceptibility phenotype is transmissible beyond the first post-radiation generation. The induced rates of tumors were 100-fold higher than those known for mouse specific locus mutations. There were clear strain differences in the types of naturally-occurring and induced tumors and most of the latter were malignant. Another important finding was that germ-line exposure elicited very weak tumorigenic responses, but caused persistent hypersensitivity in the offspring for the subsequent development of cancer by the postnatal environment. Activations of oncogenes, ras, mos, abl, etc. and mutations in tumor suppressor genes such as p53 were also detected in specific tumors in cancer-prone descendants. However, the majority of tumors observed in the progeny were those commonly observed in the strains that were used and oncogene activations were rarely observed in these tumors. It can be hypothesized that genetic instability modifies tumor occurrence in a transgenerational manner, but so far no links could be established between chromosomal and molecular changes and transmissible tumor risks. Our data are consistent with the hypothesis that cumulative changes in many normal but cancer-related genes affecting immunological, biochemical and physiological functions may slightly elevate the incidence of tumors or fasten the tumor development. This hypothesis is supported by our GeneChip analyses which showed suppression and/or over-expression of many such genes in the offspring of mice exposed to radiation. In humans, a higher risk of leukemia and birth defects has been reported in the children of fathers who had been exposed to radionuclides in the nuclear reprocessing plants and to diagnostic radiation. These findings have not been supported in the children of atomic bomb survivors in Hiroshima and Nagasaki, who were exposed to higher doses of atomic radiation. However, it will be important to follow the human subjects, especially for adult type cancers and chronic diseases throughout their lives to determine whether the mouse studies can predict human responses.