Table 2. Leukocytes (>10,640/ 1) Hemoglobin Frequency of abnormal hematologic values over 5-year study period, 1975-1979. Sex Ronge lap Ailingnae Utirik Control F M 6/28%* (24%) 9/25 (36%) 2/8 (25%) 0/5 (0%) 22/57 (39%) 12/39 (31%) 18/64 (28%) 18/50 (36%) (>17.2 g/dl) M F 0/25 7/25 (28%) (0%) 2/5 (40%) 1/8 (13%) 15/39 (38%) 0/57 (0%) 17/50 (34%) 0/64 (0%) Hemoglobin (<12.0 g/dl) F M 9/25 2/25 (36%) (8%) 0/8 0/5 (0%) (0%) 16/57 (28%) 3/39 (8%) 47/64 (73%) 6/50 (12%) Platelets (>372,000/uL) F M 7/25 2/25 (314) (8%) 4/8 (50%) 1/5 (20%) 29/57 (51%) 12/39 (31%) 20/64 (31%) 8/50 (16%) *Denominators are average numbers of individuals seen in each category on annual or semiannual field trips. the Marshallese than in Americans (11). Peripheral blood eosinophilia in the region of 5% was noted consistently in previous studies and has persisted dur- ing the past five-year interval with no clear differences among the various groups. The children of individuals in the exposed and comparison populations have also been followed hematologically with the results recorded in Table 3. Differences between sexes for mean values of platelets, erythrocytes, hemoglobin, and leukocytes were not as prominently nor as consistently seen as in their parents. No significant differences in any measurement could be related to exposure status, but the numbers of individuals in each subset were relatively small. The established association between radiation exposure and development of leukemia has focused attention on this potential in the Marshallese. In 1972, a 19-year-old, who was the youngest male exposed on Rongelap in 1954, developed and rapidly succumbed to acute myelogenous leukemia. He presented with asymptomatic leukopenia (2000/u1) and thrombocytopenia (120,000/u1) detected on the routine annual survey, and the diagnosis was quickly estab- lished by bone marrow examination. His history and clinical course have been reported in detail previously (1,33), and he represents the only instance of leukemia in this series. Statistical calculations, based on the best available data for leukemia incidences in the U.S. and Micronesia, indicate a 5- to 15-fold greater probability that this Marshallese case was radiation-induced -23-