by previous loading experiments on a similar-size structure in Upshot-Knothole Project 3.1 and high explosive tests by Sandia Corporation at the Coyote Canyon site, Sandia Base. Gages were placed in pairs at various locations on the front, top, and back of this structure; the pairing allowed determination of how closely two independent gages of the Wianko type would agree under air blast. As it developed, the Castle Project 3.1 structure was exposed to a blast from Shot 3, which had a yield (130 kt) of only about a tenth of that predicted. Thus the peak overpressure was only about 3.5 psi instead of the 12 to 15 psi predicted. Although the specific objective of the project was therefore not accomplished, it was believed that much useful infor mation could still be obtained from the data subsequent to the shot. Two blast-loading methods had been developed which could possibly be checked by this data. The blast-loading method in AFSWP~-226 had been developed by Sandia Corporation based Figure 4.1 Test cubicle, Project 3.1. Left: front view. Right: rear view. on high explosive, shock-tube, and full~scale data; the Armour Research Foundation (ARF) method was a blast~loading procedure developed by the ARF based on shock-tube and fullscale data. Consequently, an evaluation of the | last-loading data from this project was undertaken by Sandia Corporation to (1) make a comparison of the blast loading on the two Upshot-Knothole and Castle structures (which were of approximately the same dimensions) when subjected to blast waves having the same peak incident overpressure but different positive-phase duration; (2) evaluate the accuracy of oth the so-called AFSWP-226 and ARF loading-prediction procedures against the pressure loading indicated by the centerline gages of Castle Structure 3.1— since the procedure set forth in AFSWP-~226 is predominantly applicable to two-dimensional structures, the gages at the center line of the structure were expected to give the best agreement; and (3) assess the reproducibility of Wianko gage measurements from the records of gage pairs on Castle Structure 3.1. The results of this evaluation by Sandia Corporation indicated the following. The AFSWP-~226 loading-prediction procedure gave reasonably good results. Also, the agreement of both AFSWP-226 and ARF predictions (within the diffractive phase) with the centerline gage records of the two full-scale tests was reasonably good. The net-loading curves produced with both the AFSWP-226 and ARF prediction procedures (within the 59