Eighty-two high performance young male fast bowlers (mean age 16.8 years) were tested immediately prior to the season for selected kinanthropometric and physiological data. Subjects were also filmed both laterally (200 Hz) and from above (100 Hz) while bowling so that their front foot impacted a force platform during the delivery stride. The players then completed a log book over the ensuing season that detailed their training and playing programmes. All cricket related injuries over this season were assessed by a sports physician who used computerized tomography to assist in the diagnosis of spinal injuries. At the completion of this season the players were grouped according to their injury status (Group 1--bony injury to a vertebra; Group 2--soft tissue injury to the back that caused the player to miss at least one game, and Group 3--no injuries). A one-way analysis of variance was used to identify if any variables were significantly (P less than 0.05) different between the three groups, and a Scheffe post hoc comparison was used to determine which groups were significantly different. Eleven per cent of the players sustained a stress fracture to a vertebra(e) (L4 to S1), while 27 per cent sustained a soft tissue injury to the back. Bowlers with a low longitudinal foot arch were more likely to develop a stress fracture than those with a high arch. Shoulder depression and horizontal flexion strength for the preferred limb and quadriceps power in the non-preferred limb were also significantly related to back injuries. Results suggest that bowlers with the above physical characteristics, who bowl with these biomechanical techniques for extended periods, are predisposed to back injuries.