estimate the health risk that may be associated with long-term exposures at-the level of the radiation dose and soil removal criteria being recommended. It is clear from this review that knowledge of the relationship between radiation dose and effects of that dose on man as characterized in dose-effect curves is incomplete even for external radiation exposures. For internal emitters and particularly for plutonium, the situation is even less satisfactory. UNSCEAR has summarized their findings by stating that one should not extrapolate ina linear fashion from effects seen at high doses and dose rates to effects at low doses and dose rates since there is strong likelihood of recovery and repair. The BEIR Committee, using only human data, concluded that since the low dose data were incomplete, one should conservatively assume a linear no-threshold dose-effect curve drawn through data obtained at high doses and dose rates. The committee further suggested that if this linear no-threshold curve is assumed to be correct, it follows that 6,000 cases of cancer would be produced each year in a population of 200, 000, 000 people exposed ata rate of 0.17 Rem/yr. (This is the FRC RPG for population groups - see Table I.) For the Enewetak population of less than 500 exposed at the sarme level, one can makethe following estimate: 6X 10° cases/yr X 500 people = 2X 10° people 1.5X 1072 cases of cancer/yr Using a linear dose-effect curve, exposure at the level of the recommended criterion of 0,25 Rem/yr would give 2.2 X 10-2 cases per year. The Task Group views this as a pessimistic upper limit of risk. It could be inferred that there may be between zero and three cases of cancer in 100 years if the entire Enewetak population were continuously exposed to 0.25 Rem/yr over that time period. Most of the exposure to whole body, at Enewetak, and in fact, to all organs will come from internal emitters. The shape of the dose-effect curve for exposures from internal emitters is most uncertain because of lack of experience and lack of confidence in extrapolation of high dose and dose rate effects into the very low dose and low dose rate situation. A lack of confidence in JH-12