YIELD MEASUREMENTS. methods, Device yield is usually determined by several two of which involve photo-optical techniques. Growth of the in- tensely hot and radiating mass of device debris and air that constitute the nuclear fireball varies with its yield. Cameras were therefore used to record this growth, and film records subsequently analyzed to infer yield. The duration and the intensity of the energy pulse in the optical- thermal spectral region also vary with yield; thus, light detectors coupled to recorders were also used to derive yield. In addition, yield may be determined by collecting and analyzing a representative sample of the device debris. Inferences are then drawn regarding the yield, based on knowledge of the materials in the unexploded device. The construction, instrumentation placement, and data recovery for the photo-optical yield determinations did not usually require personnel to be in areas with a high potential for exposure to radiation. Cameras and light detectors need only a clear field of view of the burst point and enough breadth of view to encompass the fireball. Camera placement did not involve personnel at times and places of heavy contamination. Film recovery generally did not involve high exposure potential, as the photo stations were usually at ranges and in directions not heavily contaminated by fallout. The sampling of-device debris, however, necessitated much closer contact with higher levels of radioactivity. The technique used in CASTLE and most atmospheric tests was to fly aircraft with collectors directly through portions of the radioactive cloud. About 90 percent of the fis- sion debris was usually considered to be in the upper portion of the radioactive cloud (Reference 2). representative sample. Several aircraft were used to obtain a The aircrews were exposed to the radiation emitted by the radioactive particles in the cloud as they flew through. The air- craft flying these sampling missions picked up significant amounts of radioactive material on their surfaces, posing additional and continuing 32