was swept slowly across the pulse spectrum and the output simultaneously fed into a count-rate meter. Count rate was recorded on the Y-axis of the plotter, and the analyzer base-line position on the X-axis, giving a record reducible to gammaintensity versus energy. The detection unit consisted of a 4-inch-diameter-by-4-inch-thick Nal(T1) crystal, optically coupled to a photomultiplier tube and housed in a lead shield 2, inch thick on the sides and bottom. A 6- inch-thick lead plug witha 4 -inch-diameter collimating opening was located on top, with the collimator directed toward the center of the crystal. The sample was placed in a glass vial and suspended in a fixed position a short distance above the collimator. All quantitative gammaenergy-spectra measurements of cloud and fallout samples were made with this instrument. Relative spectral data was also obtained at later times with a single-channel analyzer. This instrument utilized a detection unit with a 3-inch-diameter-by-3-inch-thick uncollimated Nal(T1) crystal. Reproducible geometries were neither required nor obtained; energy calibration was accomplished with convenient known standards. 2.3. STATION LOCATIONS 2.3.1 Barges, Rafts, Islands, and Skiffs. The approximate locations of all project stations in the atoll area are shown for each shot in Figure 2.10; more exact locations are tabulated in Table 2.3. The Rafts 1, 2, and 3, the island stations on Sites George and How, and the Skiffs DD, EE, KK, LL, and TT remained in the same locations during the entire operation. Other Stations changed position at least once and sometimes for each shot. These changes are indi- cated on the map by the letters for the shots during which the given position applies; the table, however, gives the exact locations. All stations were secured and protected from fallout during Shot Dakota in which this project did not participate. The choice of locations for the barges was conditioned by the availability of cleared anchoring sites, the necessity of avoiding serious blast damage, and the fact that the YFNB 29 carried two major sampling arrays while the YFNB 13 carried only one. Within these limitations they were arranged to sample the heaviest fallout predicted for the lagoon area and yet guard against late changes in wind direction. In general, the YFNB 29 was located near Site How for all shots except Tewa, when it was anchored off Site Bravo. The YFNB 13 was located near Site Charlie for all shots except Cherokee and Tewa, when it was positioned near Site How. Because both barges were observedto oscillate slowly almost completely around their points of anchorage, an uncertainty of +200 yards must be associated with the locations given in Table 2.3. The raft positions were chosen for much the same reasons as for the barge positions, but also to improve the spacing of data points in the lagoon. An uncertainty of +150 yards should be associated with these anchorage coordinates. The island stations, except for Site How, were selected on the basis of predicted heavy fallout. It was for this reason that the minor sampling array (M) located at Site William for Shots Cherokee, Zuni, and Flathead was moved to Site Charlie for Shots Navajo and Tewa. Site How was selected to be in a region of moderate fallout so that survey and recovery teams could enter at early times. A detailed layout of the installation on Site How is shown in Figure 2.8. Because the skiffs were deep anchored and could not be easily moved (Reference 15), their locations were originally selected to provide roughly uniform coverage of the most probable fallout sector. With the exception of Stations WW, XX, and YY—assembled from components recovered from other stations and placed late in the operation—their positions were not delib- erately changed. Instead, the different locations shown in Figure 2.10 reflect the fact that the skiffs sometimes moved their anchorages and sometimes broke loose entirely and were temporarily lost. Loran fixes were taken during arming and recovery, before and after each shot. The locations given in Table 2.3 were derived from the fixes and represent the best estimate of the positions of the skiffs during fallout, for an average deviation of + 1,000 yards in each coordinate. 2.3.2 Ships. The approximate locations of the three project ships at the times when they ex- perienced peak ionization rates during each shot are presented in Figure 2.11. 24 Table 2.4 gives