detonation. This dust would be large sized, 320 average diameter, and be visi- ble on food preparation surfaces. 5. Activity Intake by Breathing Contaminated Air Mean air concentration estimates of the activity of selected nu- clides were based on the deposition rates of fallout. The percent of activity deposited per minute at a point in time at Rongelap Island was estimated from the slopes of tangents to the curve in Fig. 4. point in time was estimated from Eq. 2. The diameter of particles at a The velocity corresponding to particle diameter was taken from Fig. 8 of Holland (Ho63). The air activity concentration at a point~in time was assumed, 1) directly proportional to the percent of total activity deposited per minute at that point in time, 2) directly proportional to the activity on the ground at the end of fallout corrected for decay back to that point in time, and 3) inversly proportional to velocity of fall of granules at that time. This same method was applied to surface activity data for Sifo and Utirik Islands as well. Values for air concentration at all three islands and times post BRAVO detonation were tabulated and given in Table 10. The air activity concentrations for Rongelap Island derived from Bikini Ash data were in agreement with air activity concentration data derived by Peterson (Pe81). Peterson used the MATHEW-ADPIC code suite and all the observed wind data (see Section II). The air activity concentrations for Utirik Island relative to those at Rongelap Island might be expected to be less because of the exposure rate differences that were observed. The velocity of fall of a granule corresponding to the activity median diameter was greater by a factor of 95 at Rongelap Island when compared to Utirik Island, while the exposure rate after deposition differed by only a factor of 9.5. 32 If the deposition intervals at