a fraction of one per cent of the fallout deposited in 1954. Therefore the Redwing and Hardtack fallout are insignificant when considering the long-range picture. Selection of areas and reproducibility of repeated measurements In March 1958, stakes were set out in various parts of Rongelap islet in an attempt to provide a means of repeating measurements at identical locations, Stakes proved to be unsatisfactory since they only served to attract the curious, which resulted in trampling and disturbance of the areas and in some cases removal of the stakes. A practical solution to this problem was to select general areas within which measurements were to be taken. These areas were located with relation to pathways, roads, buildings, and measured distances from landmarks. In each general area measurements were taken over different types of vegetation, soil, the pathways themselves, over litter, and under Pandanus trees and other tall plants. In August 1958, the set of measurements was repeated three times at Rongelap islet and twice at both Eniaetok and Kabelle islets. At Rongelap islet the average for each set of readings ranged from 0.046 to 0.067 mr/hr. At Eniaetok the range of the aver- age readings was 0.073 to 0.079 mr/hr and at Kabelle islet, ed as 0.137 to 0.178 mr/hr. UND Gb ow, STAIRS