fallout studies at the Nevada Test Site (Hi81). An approximation of the exposure due to all not-accounted-for nuclides listed above would be less than 12 of the total from day 0.5 to day 2.2 post-detonation. This may not be a good comparative assessment since no study of a device like BRAVO was reported at the Nevada Test Site. The fourth shot in the Castle Series, UNION, was a 9.5-MT device with a number of characteristics similar to those of BRAVO, including its fission-to-fusion ratio. The gross gamma decay was measured from 0.2 day post~detonation, unlike BRAVO for which gross gamma decay was not reported until after 5.5 days had passed. From 0.2 day to 0.3 day the decay exponent was -1.5. From 0.2 day to 2.2 days it was -1.3. From 2.2 days to 26 days it was -1.4. These values of gamma-decay exponents were comparable to those we estimated from Bikini ash extrapolations. Significant amounts of not-accounted-for nuclides in the activity produced by UNION might have caused us to estimate a wide difference in decay exponents. Thus, we conclude that not-accounted-for nuclides contributed very little to the dose received by persons at Rongelap, Utirik, or Sifo Islands. v. Input Data to Kinetics Equations. A check on input data for activity per unit fission versus time was made by us. The theoretical activity of unfractionated iodine isotopes following 10,000 thermonuclear fissions of U as given by Crocker (Cr65) was compared to the activity at any time follow- ing fission of U with 14-MeV neutrons. Our comparison calculation was based on decay schemes from the Table of the Isotopes (Le78), independent yield data for fission products from the National Nuclear Data Center (NNDC82), and Eq. (4). The Crocker yields were based on a slightly different neutron energy spectrum than that used in the calculation made here. The kinetics equations, verified yield data, and our decay scheme approach led to results remarkably similar to those of Crocker. Our estimated maximum difference was approximately 50% for 1341 activity at two hours post-detonation. All iodine isotope activities were within 20% of the Crocker estimate at about 10 hours postdetonation. vi. Exposure Rate on Day 0.5 and Surface Roughness Effects. We estimated the mean glanr rate from all the nuclides given in Table 13 to be 7.9 x 10°77 c kg! s7! (1.1 x 10! R n7!) on day 0.5 post-detonation. in this estimate the contribution from noble gas nuclides; We include Exclusion of the noble gas activity yields 7.0 x 1077 C kg7! s7! (9.7 R h7!) for the exposure rate at day 0.5. The exposure-rate history at Rongelap Island based on the particulate activity in Bikini Ash, has been plotted in Figure 10. Photons emitted by fallout on the surface may have been intercepted by overlying layers of soil or by surface structures. Because of surface roughness effects, the radsafe survey team may have reported an exposure rate which is an underestimate of one which is produced by a plane source. This would have caused us to underestimate surface activity at Rongelap because Bikini ash activity was normalized to the Rongelap exposure-rate survey made by the radsafe team. The effect of fallout particles penetrating into Rongelap's coral surface was approximated by comparing Beck's values for mR h7! uci7! m for activity on a smooth flat plane, toa plane _yhere activity was distributed depthwise with a relaxation length of 0.16 g cm”2 (Be80). By relaxation length we mean the depth at which the activity in overlying layers of soil is reduced - 38 -