One-hundred-and-two high-level players of the field-games soccer, Gaelic football and hurling began a two-year investigation into the intrinsic causes of sports-injuries; 86 completed the study. During the first year all injuries, and the time affected by injury, were recorded. The subjects then underwent flexibility tests, an accurate photogrammetric assessment of posture, measures of speed and acceleration, and a clinical assessment of anatomical and physiological factors thought to be associated with the risk of sports injury. Time affected by injury was then recorded for a further 12-month period. Stepwise multiple-regression analysis revealed that the number of days of injury during the second 12-month period could be predicted from (1) the days of injury during the first 12-month period, (2) posture, (3) acceleration over 10m from a standing start, and (4) the number of musculo-skeletal clinical defects. Flexibility scores were not found to be significant predictors of injury. It is suggested that injury prevention programmes should concentrate on improving posture and the rehabilitation from previous injury rather than flexibility; and that research should be undertaken into the effectiveness of such interventions.