HEPATITIS B SEROLOGY IN MARSHALLESE 77 surface antigen-positive individuals were 187 + 48 for the Rongelap exposed (N = 2), 177 + 40 for the Utirik exposed (N = 21), and 179 + 40 for the comparison group (N = 10). These values do notdiffer significantly (P > 0.80). Of 34 HBsAg-positive per- sons detected, 8 (23.5%) were also positive for e antigen, a correlate ofinfectivity. Antibodyagainst 6 agent was not detected in any of the Marshallese tested. DISCUSSION It is now 31 years since radioactive fallout settled on Rongelap and Utirik. The number of Marshallese remaining in the radiation-exposed groups is small in com- parison with the population ofJapanese atomic bomb survivors being followed by the Radiation Effects Research Foundation. Nevertheless the high prevalenceofserologic markers of hepatitis B infection in the Marshallese permits statistical inferences of differences amongthe three exposure groups. This analysis indicates (i) persons from Utink, who suffered a relatively low radiation exposure, have the highest prevalence of HBsAg;(ii) persons from Rongelap whoreceived a high exposure have the lowest prevalence of HBsAg;(iii) unexposed persons living on Rongelap also have a relatively low frequency of HBsAg; and(iv) the prevalence ofat least one marker for HB was similarin all three groups. There is reason for concern about a possible clinically significant depression in immune competence in the exposed people of Rongelap. Peripheral blood lymphocyte responsiveness to phytohemagglutinin has been reported as diminished in the heavily exposed Japanese population (1/1). Complementfixation tests performed in 1959 for several infectious diseases revealed most meanantibodytiters to be lower in the Rongelap exposed than in the comparison population (12). There was, in addition, a lower primary antibody response to tetanus toxoid among the exposed Rongelap group when challenged 3 years after the fallout, although the difference between exposed and unexposed was notstatistically significant (/3). Despite these findings, immunoglobulin levels, phytohemagglutinin responsiveness, and tuber- culin and Candida skin test responsiveness have been similar in the two groups (9, 14, 15). There has been no documentedincrease in any type of infection noted as a delayed consequence of radiation exposure in either the exposed Japanese or the Marshallese. With regard to hepatitis B infection, HBsAg titers among the Japanese exposed to > 100 rad are similar to those who received <10 rad, and the occurrenceof antibodies to HBsAgis similar in the high- and low-dose groups (7). Nevertheless, studies by the Radiation Effects Research Foundation have revealed a suggestive association between radiation dose and prevalence of cirrhosis in atomic bomb survivors, and the prevalence of HBsAgis 3.4% in the high-dose and 2.0% in the low-dose groups, a statistically significant difference (6, 7). These investigations have suggested therefore that HBsAgpersistence, but not susceptibility to the infection itself, might be the consequence of immunologic impairmentin the high-dose group. With this backgroundit is reassuring to note that the prevalence of HBsAg was relatively low among the exposed Rongelap group rather than elevated.It is unlikely Ci ~ ara i} co that this finding is due to earlier attrition among HBsAg-positive exposed persons