A series of fracture tests on large-scale, pre-cracked, aluminum alloy panels was carried out to examine and to characterize the process by which cracks propagate and link up in this material. Extended grips and test fixtures were specially designed to enable the panel specimens to be loaded in tension in a 1780-kN-capacity universal testing machine. Ten panel specimens, each consisting of a single sheet of bare 2024-T3 aluminum alloy, approximately 4 m high, 2.3 m wide, and 1 mm thick, were fabricated with simulated through-cracks oriented horizontally at mid-height. Using existing information, a test matrix was set up to explore regions of failure controlled by failure mechanics, with additional tests near the boundary between plastic collapse and fracture. In addition, a variety of multiple site damage (MSD) configurations were included to distinguish between various proposed linkage mechanisms. All tests but one used anti-buckling guides. Three specimens were fabricated with a single central crack, six others had multiple cracks on each side of the central crack, and one had a single crack but no anti-buckling guides. The results of each fracture event were recorded on various media: film, video, computer, magnetic tape, and occasionally optical microscopy. The video showed the crack tip with a load meter in the field of view, using motion picture film for one tip and super VHS video tape for the other. The computer recorded the output of the testing machine load cell, the stroke, and the twelve strain gages at 1.5-s intervals. A wideband FM magnetic tape recorder was used to record data from the same sources. The data were analyzed by two different procedures: (1) the plastic zone model based on the residual strength diagram, and (2) the R-curve. The first three tests were used to determine the basic material properties, and these results were then used in the analysis of the subsequent tests with MSD cracks. There is fairly good agreement between measured values and results obtained from the models.