Theactivities of the IC trays have not been adjusted for sampling bias, although some undoubtedly exists, primarily because its quantitative effects are unknown. Relative rates may still be derived if it is assumed that all trays are biased alike, which appears reasonable for those cases in which wind speed and direction were nearly constant during the sampling period (Section 4.3.2). More extensive analysis would be required to eliminate uncertainties in the remaining cases. It should also be mentioned that IC trays with alternating greased-disk and reagent-film collecting surfaces were intentionally used in all of the collectors for Shots Flathead and Navajo — with no detectable difference in efficiency for the resulting fallout drops—and of necessity for Shot Tewa. The late move of Shot Tewa to shaliow water produced essentially solid particle fallout, for which the efficiency of the reagent film as a collector was markedly low. Thus, only the greased-disk results have been plotted for the YAG 40 in Figure 3.4, although it was necessary to plot both types for some of the other stations. Trays containing reagent-film disks, all of which were assigned numbers between 2994 and 3933, may be distinguished by reference to Table B.2. A few trays, designated by the prefix P, also contained polyethylene disks to facilitate sample recovery. 3.2.2 Times of Arrival, Peak Activity, and Cessation. The times at which fallout first arrived, reached its peak, and ceased at each major array station are summarized for all shots in Table 3.1. Peak ionization rates are also listed for convenient reference. Time of arrival detector (TOAD) results, covering all minor array stations and providing additional values for the major stations in the atoll area, are tabulated in Table 3.2. The values given in Table 3.1 were derived from Figures 3.1 through 3.4, and the associated numerical values in Tables B.1 and B.2, by establishing certain criteria which could be applied throughout. These are stated in the table heading; while not the only ones possible, they were felt to be the most reasonable in view of the available data. Arrival times (tg) were determined by inspection of both TIR and IC records, the resulting values being commensurate with both. Because the arrival characteristics varied, arrival could not be defined in some simple way, such as “1 mr/br above background.” The final values, therefore, were chosen as sensible-arrival times, treating each case individually. It should be mentioned that, within the resolving power of the instruments used, no time difference existed between the onset of material collections on the IC trays and the toe of the TIR buildup curve. The IC’s on the ships were manually operated and generally were not triggered until the arrival of fallout was indicated by the TIR or a survey meter, thus precluding any arriva] determination by IC; those at the unmanned stations, however, triggered automatically at shot time, or shortly thereafter, and could be used. The SIC on the YAG 40 also provided usable data, ordinarily yielding an earlier arrival time than IC B-7 on the same ship. In order to conserve trays, however, the number exposed before fallout arrival was kept small, resulting in a larger time uncertainty within the exposure interval of the first active tray. Once defined, times of peak activity (tp) could be taken directly from the TIR curves. Because peaks were sometimes broad and flat, however, it was felt to be desirable to show also the time interval during which the ionization rate was within 10 percent of the peak value. Examination of these data indicated that tp ~2tg ; this point is discussed and additional data are presented in Reference 18. Cessation time (t,) is even more difficult to define than arrival time. In almost every case, for example, fallout was still being deposited at a very low rate on the YAG 40 when the ship departed station. Nevertheless, an extrapolated cessation time which was too late would give an erroneous impression, because 90 or 95 percent of the fallout was down hours earlier. For this reason, IC-tray activities measured at a common time were cumulated and the time at which 95 percent of the fallout had been deposited read off. A typical curve rises abruptly, rounds over, and approaches the total amount of fallout asymptotically. Extrapolated cessation times were estimated primarily from the direct IC plots (Figures 3.1 through 3.4), supplemented by the cumulative plots, and the TIR records replotted on log-log paper. It must be emphasized 44