60 conversion factor for each sample (density) was then approximated by interpolation, assuming that the actual conversion factor varied linearly with density between the range of factors determined in the 1.17 and 1.35 g/cc standards. The error that could result due to a nonadherence to the linear dependence assumption described above was estimted by considering the case where density changes give rise to logaritnmic rather than linear changes in the correction factor. The maximum error that could result from a logarithmic instead of the assuved linear dependence was estimated by finding the difference in the value cf the two correction factors which occurred at the extremes of sample density encountered in this work (0.6 and 1.6 g/cc). The difference in correction fectors thus determined, using the two different correction factors, was 7.3% fcr the sample geometry and density limit which yielded the highest error when counting 60 KeV gamma rays. For radionuclide concentrations which were deter-~ rined by using higher energy gamma-rays, and for the majority of samptes which were not at the extreme limts of densities, the error which could arise due to this uncertainty is smaller than 7.3%. The abundance of each y-ray observed for a radionuclide (Table 11) was used to calculate the concentration of the radionuclide present. Where more than one radiation from one nuclide was observed, the reported value is that cerived by weighting the concentrations determined from each gamma peak observed by its associated relative counting error, and a weighted mean concentration and error was determined (Stevenson, 1966). The error term associated with the counting of individual energy gamma-rays are 2 S.D. errors based on propagated counting statistics. The concentrations of all the isotopes measured by y-spectrometry in this work are corrected for decay to the date of the collection shown in Tables 3 and 4,