PROCEDURES Gamma ray spectrum analysis and Sr analysis on all coconut, pandanus, and arrowroot samples was performed at the SWRHL Pacific Operations Laboratory in Honolulu, Hawaii. Edible portions of the samples were ground and counted without drying on a 4" x 4" NaI (T1) crystal multi-channel analyzer system. The spectral range covered was zero to two MeV. At the concentra- tions of !37Cs encountered in these samples the counting error is approximately ten percent. (All errors referenced in this section are 2 sigma.) Following gamma spectrum analysis the samples were ashed and the inorganic residue was analyzed for 9°Sr. Due to the small amount of ash produced it was often necessary to composite samples from adjacent locations. Counting was performed on a Jow background beta counter. The analytical error associated with this procedure is approximately ten percent at the levels of 99Sr encountered. All tritium and plutonium analyses were performed at SWRHL, Las Vegas, Nevada. Tritium was determined by liquid scintillation counting of the water recovered from coconut milk and from selected soil samples. The minimum sensitivity for this procedure is €.4 pCi/ml for five ml of recovered water. At this concentration the error is 100 percent. Analyses of the coconut milk and soil samples indicated the concentrations of tritium were Tess than the minimum detectable level. Plutonium analyses, which were performed only on soil samples, were by radiochemical separation followed by pulse height analysis using a lithium drifted silicon detector. The analytical error is approxi- mately ten to fifteen percent. Unless otherwise specified each soil sample represented the top one to two inches of a one-square-foot area at each location. Prior to plutonium analysis, gamma ray spectrum analysis was performed to identify the most abundant radionuclides and to estimate their relative contribution to the total activity.