- 50 Saturated sodium carbonate were added to 100 cc of sample, the sample then filtered through millipore filter paper, and the precipitate ashed and counted. The purpose of the sodium carbonate was to remove naturally occurring «40 (approximately 540 d/m/liter) from the precipitate. Tests performed by the Health and Safety Laboratory, New York Operations Office, on this technique indicate that potassium does not co-precipitate and that from spikes using 18-month-old mixed fission products , only 23 percent of the activity escaped precipitation!®, The counting error per unit weight is considerably less in the "treated" column because of larger sample size and a smaller correction for self-absorption. Values in the "untreated" column for the October collection indicate some radioactivity in the Kabelle sample, but a questionable amount of activity in the three other samples. However, the values in the treated” column definitely indicate activity in both the Kabelle and Labaredj samples. The best estimate of the maximum value of mixed fission product activity in Rongelap lagoon water is 0. 0006 yc per liter. The island water samples were prepared for counting by drying both an unfiltered and filtered sample. The filtered sample was passed through a millipore filter and both the filtrate and the residue were counted. One hundrecd-cc samples were used without difficulty as there was very little residue remaining on the plates after drying. However, the well sample from Rongelap Island had some salts and from the weight of the salts it was estimated that 8 percent of the sample was sea water. The well water on Rongelap Island was less radioactive thm cistern water.