The clinical mechanical failures of small diameter intramedullary interlocking nails were evaluated to determine the relationship of failure modes to the type or location of tibial fractures. Methods were developed to duplicate failure modes in vitro in standardized tests to simulate the clinical situations. Where standard test methods were inadequate, new methods were developed to provide quantifiable, reliable methods of evaluating potential clinical performance. The modes and rates of mechanical failure in the clinical series were consistent among participating centers: (1) In diaphyseal fractures with secondary trauma, the intramedullary nail bent at the fracture site where the working length was unsupported; (2) failures that occurred several weeks after nailing were the result of fatigue fractures of the locking screws, usually at the distal end; and (3) nail and screw failures occurred most commonly in proximal and distal tibial fractures. The strength of the 8- and 9-mm sizes of Synthes and Russell-Taylor nails were comparable.