islands where °°co and *#*°Sb are now the primary contributors, ee ara wt ATW Ge teen amp RE RSME CEE Er etetet the 1964 soil data usually indicate that these two isctopes contributed even a larger percentage of the total exposure in 1964; often even larger than would be estimated from half life alone, suggesting that weathering may play an important role on these islands. The concentrations of various radioisotopes in the soil varied considerably from island tc isiand with the islands further from detonation sites exhibiting mostly fission product activity, while islands close-in to detonation sites exhibited a variety of both fissicn and activation products. The particular y-ray emitters found on each island and their relative concentrations are discussed further in Section Iff which treats in detail the environmental radiation fields on each individual island. Error Estimates and Data Evaluation te Cc. Error in Total Gamma-Ray Exposure Rates ~- Based on our past experience with these instruments, the consistency of the field data, and the laboratory calibrations. total exposure rates at specific sites surveyed with the ionization chamber and/or the spectrometer system are estimated to be accurate to within 1 ur/hr. Total exposure rates obtained with the portable scintillation detector are probably about + 10% S.D. and those with the G-M survey meters about + 20% S.D. The overall accuracy in the range of the measurements for the sites surveyed is probably better than 10%. a ent Error in Partial Exposure Rates for Major Emitters - The partial exposure rates obtained from the field spectrometric measurements tor *2" cs, accurate to + 10% S.D. sity 1e >ion vas was =" Co, and “°*°Sb are estimated to be for the first two emitters and + 20% S.D. for the latter. The ~#" Sb estimate sometimes includes a small contribution from ~~°™Rh. This estimate is based on the calibration accuracy as well as the amount by which the sum of the individual exposure rates differs from the independently measured total exposure rate. In most cases this difference was less than 10% and is probably due to errors in the assumed depth distribution relaxation length, non-uniformity of the radiation field, ground