We have studied frequencies of mutations induced at the b locus of the fish, Medaka Oryzias latipes, after gamma-irradiation. Homozygotes for the b locus have colorless melanophores whose phenotypic expression can be distinguished from that of the wild type. An advantage of the use of oviparous fish for detection of skin color mutations is that the mutant phenotype can be confirmed as early as 1.5 days after fertilization because of the transparent egg membrane of the embryo. Wild-type (B/B) male fish were exposed to 4.75 or 9.5 Gy of 137Cs gamma-rays at a dose rate of 0.95 Gy/min and then mated with the female testers (b/b). A total of 77,761 F1 offspring were examined for mutation and other abnormalities. In the control, we had 1 mutant among 22,068 offspring, resulting in a mutation rate of 4.53 X 10(-5)/locus/gamete. However, this mutant embryo died before hatching. Therefore, in an attempt to present specific-locus mutation frequencies in the fish, the frequencies of color mutants that survived more than 4 days after hatching were used as frequencies of viable mutants; (number of viable color mutants)/(number of hatched fry that survived more than 4 days after hatching). In the 4.75 Gy-irradiated group the viable mutant frequencies were 45.0 X 10(-5), 69.7 X 10(-5) and 0/locus/gamete, while exposure to 9.5 Gy resulted in mutation rates of 217 X 10(-5), 130 X 10(-5) and 8.06 X 10(-5), respectively, for sperm, spermatids and spermatogonia. In comparison with viable color mutant frequencies those of the total color mutants, which include such mutants as ones that died before hatching (defined as number of total color mutants/number of fertilized eggs minus number of early deaths), were considerably higher. For sperm, spermatids, and spermatogonia after exposure to 4.75 Gy, the frequencies were 1180 X 10(-5), 629 X 10(-5) and 9.90 X 10(-5)/locus/gamete, respectively, and in 9.5-Gy-irradiated fish, the frequencies were 1940 X 10(-5), 953 X 10(-5) and 55.5 X 10(-5). Although our data are incomplete, the present results were compared with mutation induction in mice. We concluded that the frequencies of viable color mutants in the fish can be compared with those in mice.