infections or chemicals (14, 16, 17). However, such factors are probably not involved since they did not produce a similar effect in the irradiated group who were exposed to the same factors. Correlation of chromosomal aberrations and severity of early radiation syndrome in the Marshallese was not apparent. For instance, the degree of early hematological depression did not seem to correlate with the occurrence or number of chromosomal aberrations on an individual basis. And, as was pointed out, the lower dose group had a higher incidence of chromosomal aberrations than did the higher dose group, though the radiation effects were more pronounced in the latter. These findings are in conformity with those of the ABCC studies (12) which failed to show correlation of chromosomal aberrations with severity of early radiation illness. There was also no correlation of chromosomal findings in the Marshallese with age or sex. Neither did the presence or absence of thyroid pathology appear to be related. However, the development of thyroid abnormalities was related to the dose to the thyroid gland from internally absorbed radioiodines (based largely on the size of the gland) rather than the whole-body exposure. Similar chromosome aberrations have been reported in the Japanese fishermen exposed to radiation from the same fallout (11). The incidence of aberrations, excluding aneuploid cells, was 2.1 per cent and thus identical with that seen in the Marshallese people. The majority were two-hit aberrations. The incidence of acentric fragments in a control population of Japanese studied by the same authors was 0.11 per cent; this is twenty times less than that seen among the eight unexposed Marshallese in the present study. In a carefully controlled cytogenetic study of a sample of survivors of the atomic bombings in Hiroshima and Nagasaki, Bloom et al.have found an incidence of chromosome aberrations of 0.6 per cent in 94 persons 20 years after exposure (12). Thirty-eight per cent of the aberrations were of the multiple-hit type and the rest acentric fragments. Only a single minute fragment was seen in 8847 cells of the 94 controls, an unusually low incidence indeed. The low number of aberrations and the fact that only 35 per cent of the survivors showed aberrations is not surprising since 20 years have elapsed since exposure. Although the Marshallese and Japanese exposed population and the conditions of their exposures are not strictly comparable, it is interesting that in the Marshallese people who were examined 10 years after exposure, the aberration rate as well as the number of aberration-positive individuals (53 per cent) was somewhat higher, as might have been expected. Acknowledgements We are greatly indebted for expert technical assistance to V. Estey, R. Hammerstrom, I. Irwin, W. Merrill and W. Waithe. We wish to thank Dr. Shields Warren for his interest in this work. We are most grateful to 140