In order to study the relationship between physical fitness/sport-specific skills and sports injuries 136 physical education students were studied during their 4-years of training in a prospective investigation. Physical fitness was measured every year using a battery of fitness tests, and the performance marks of a number of sports scored at the exams of the academy were used as parameters for the sport-specific skills. Sports injuries were recorded every 3 weeks on standard forms. Relative risk ratios were calculated between the tertile groups good, average and poor for all variables of physical fitness and sport-specific skills. Injury-proneness was defined for all and for acute and chronic injuries separately near the median number of injuries sustained. In only 6 out of 126 computed relative risks was a significant difference found. Discriminant analysis revealed an explanation of 16%, 14% and 11% of the variance for respectively all, acute and chronic injuries, at which 5 or 6 variables in varying combination were included. From our findings it may be concluded that physical fitness and sport-specific skills have little impact on sports injuries for the following two main reasons. Firstly, subjects at risk for sports injuries participate per definition in sports activities and have consequently developed their fitness and skills compared to the sedentary population. Thus, the range in physical fitness or sports skills in the population at risk is relatively small (physical education students belong to the 7th-10th decile in fitness test scores within a general college student population) and therefore an effect is hard to show. Secondly, the total number of sports injuries is very small and moreover, it should be distributed over several categories for analysis. The favourable advantages of using physical education students to study intrinsic risk factors (comparable and varied sports program, excellent compliance) appeared to be insufficient to compensate for drawbacks of selection.