The information obtained from these interviews is presented in Table 2. Table 3 lists all individuals whose 137cs body burden exceeded 75% of the maximum observed 137cg body burden in 1981. From the interviews it was determined that individuals traveled to Enjebi Island usually once per month, ate coconut meat and drank coconut milk from the LLNL garden. The trips, usually two to three days in length, were made to collect birds and eggs and were made by members of the population with an age distribution as listed in Table 4. Food from the LLNL garden was consumed during the visit and occasionally coconuts were gathered and brought back to the southern islands. While the absolute quantities of food consumed on each trip, as listed in Table 3, are subject to substantial variation, these estimates may be helpful in determining reasonable upper and lower limits of consumption for coconut meat and milk. The Marshallese were advised in the closeout meeting that a trip to Enjebi to collect birds and eggs was an acceptable practice but .consumption of food products grown in the LLNL garden would increase their 137¢5 body burden. They were further informed that this exposure to radiation did not present a health problem but the loss of data would hamper the LLNL efforts to study the environment of the northern islands. Since this would affect future use of the northern islands, the Marshallese promised to refrain from eating LLNL garden food products. Information provided during the private interviews led to the collection of three coconut samples from the LLNL garden. Gamma spectroscopy results conducted on the entire coconut (husk, shell, meat and fluid) are reported in Table 5. analysis. These coconuts have been shipped to Bill Robison for detailed If these 137¢g activity concentrations are representative of future coconut activity concentrations, then one could expect to observe 137¢g body burden of 4-7 pCi for individuals ingesting the Robison diet and residing on Enjebi Island. ‘ Table 6 presents quality assurance replicate results. Identification humbers with an asterisk indicate that the replicate count was not performed on the same whole-body counting system as the first count. The means and standard deviations reported at the bottom of the page represent results for the total program and results grouped by the method of replicate counting. The average capability to reproduce a body burden with either whole-body counting system is + 7%. The 2 sigma counting error associated with most results in Table 6 is + 5-102. Replicate counting results from the same