[en] It is extremely unlikely that a pressure tube and its calandria tube would rupture during normal operation in a CANDU (CANada Deuterium Uranium) reactor. However, if this unlikely scenario did happen, it would result in the pressure tube containing an axial, through-wall crack while still filled with water at 250-300^oC. This crack could run in the axial direction until it stopped, or it could turn and run in the circumferential direction, possibly causing a guillotine failure. As the path of this crack controls the resulting damage to the reactor, instrumented small-scale burst tests were done to determine the parameters controlling guillotine failure. These tests were analysed using the dynamic finite element code, VEC/DYNA3D. There was reasonable agreement between the measured and predicted deformation and the time and location of the guillotine failure. (Author)