ree, é “ assumed to have the same shape as that of Bikini Ash as given by Suito (Su56). Using the assumption that the activity of a granule was proportional to the 3.5 power of the radius of the granule (La65), an activity versus granule diameter distribution was estimated for Rongelap Island (see Fig. 3). The activity median diameter for fallout at Rongelap Island was about 150 microns. The activ- ity median diameter was the granule diameter corresponding to the median activity. The cumulative per cent of total activity deposited on the surface of Rongelap Island versus time post detonation (see Fig. 4) was estimated from the times of arrival and cessation of fallout, Eq. 2 and Fig. 3. The first analysis of arrival time of BRAVO fallout at Utirik Is- land, Utirik Atoll was done by Sondhaus, Sharp, Bond and Cronkite (Cr56). was estimated by them to be 22 hours post detonation. It The estimate was based on an assumed mean wind speed and the distance between Namu and Utirik Islands. Visible fallout was not observed on Utirik Island, therefore arrival and cessation times were not observed first hand. Fallout cessation was estimated by Sondhaus to be 34 hours post detonation (Cr56). New values of fallout arrival and cessation at Utirik Island were estimated based on observations on the fishing vessel, Rongelap Island and Rongerik Atoll, a military outpost. Fallout was first seen at 150 km at 3 hours post detonation by the Japanese fishermen and lasted for 5 hours (Ts55). It was then seen at Rongelap Island at 210 km at 5 hours post detonation and was reported to last for 7 hours (Sh57). Fallout was first observed at Eniwetak Is- land, Rongerik Atoll at about 7 hours post detonation (Sh57). This last value came from military personnel stationed at Eniwetak Island, 270 km from ground zero. Fallout duration at Eniwetak was reported to last into the night and per- haps into the next day (Sh57). A linear regression curve fit of the values for 12