tion was observed, a period of greatly reduced radiation intensity is clearly indicated. To emphasize this reduced rate, the followng criteria were used. When the peak dose rate exceeded the next plotted point by a factor of 10 and the interval between the two points was greater than 10 seconds, a minimum dose rate Ry, was defined for the interval by: R? Tr = —_ where R;y is the dose rate indicated by the plotted point terminating the interval and Rok is the peak dose rate starting the interval. A point on this minimum dose rate line was then selected so that the area under the figure formed by connecting the peak dose rate, the selected point, and the terminating dose rate was the same as thit obtained when the terminating dose rate was assumed constant over the whole interval. Simple geometric considerations demonstrate that such a point is uniquely determined. This treatment is admittedly arbitrary, but it at least approximates the true shape of the dose rate function more closely than the straight line connection, which is obviously in error. Although such treatment is also warranted for the other initial peaks reported, it has not been applied since both the peak rates and the time intervals involved were small enough so that the additional refinement appeared unnecessary. The early gamma records presented in Figures 3.5 through 3.31 are in excellent general agreement. No correction has been made for deposit doSe, since this correction may be safely ignored (Section 3.3.1). The records show a number ofinitial peaks followed by a period of essentially no radiation and then by a rapid increase to peak dose rate. The first part of this increase is always steeper than the latter part. This latter, more gradual rise is undoubtedly due to the approach of the base surge and its subsequent envelopment of the detector. By superimposing these early gamma records, it is generally apparent that a similar series of events occurs during both shots, the Umbrella sequence being about 10 to 20 seconds earlier. The records obtained from different instruments at the same station show some intereSting differences that are attributed to variations in detector response. The difference between the ASEL- and std-GITR records at Stations CL 4.6 and CR 4.1 on Wahoo and at Station DRR 3.9 on Umbrella are of particular interest (Figures 3.7, 3.9, and 3.23). These records showthat, although both instruments record nearly the same peak dose rate, the rise in dose rate record- ed by the ASEL-GITR always lags behind that of the std-GITR, this effect being greatest for CL 4.6 and least for DRR 3.9. In the confusion of the emergency re-arming for Wahoo, the ASEL detectors were erroneously oriented so that surface zero subtended an area of low direc- tional response (Figures C.5 and C.7; direction of surface zero in these figures is 180° and 0° for detectors at CL 4.6 and CR 4.1, respectively). These orientations are confirmed by photographs taken during instrument recovery after Wahoo. Inthe case of DRR 3.9 for Umbrella, the coracle was so positioned by the wind that an area of low response was directed toward the hot line (Appendix F and Figures C.5 and C.7; direction of hot line in these figures is 0°). Therefore, it is possible that the differences between the ASEL- and std-GITR records are the result of differences in directional response made evident by the approach of base surge. An application of the calculated detector responses (Figures 3.1 through 3.3) for the total angle subtended by an approaching base surge brings the two instruments into closer agreement, but complete agreement cannot be achieved until surge dimensions of 300 feet are assumed. This type of hypothetical approach suffers from the fact, inherent in the mathematical model, that the dimensions of the assumed radiating cloud varies as a function of distance to the detec- tor unless it is assumed to lag a certain distance behind the visible base surge boundary. This lag distance ts approximately 1,000 feet in all three cases, a distance which is in accord with other observations discussed more fully in Section 3.3.2. Photographic records indicate much larger maximum dimensions for the visible base surge; however, there are many indications (Sections 3.3.2, 3.3.4, and 3.4.1) that radioactive material is not uniformly distributed throughout the visible surge. No particular emphasis is placed on these speculations except to note that highly radioactive clouds of small dimensions are not impossible, and that these and other 85