Three hundred fifty male recruits were randomly allocated to either the standard recruit training program (N = 180) or substituted a weighted march activity for all running periods in the physical training program (N = 170). There were no other differences in the formal training program. The incidence of injury was 37.6 and 46.6% in the walk and run groups, respectively. The rate of injury was 52.9/100 recruits in the walk group and 61.7/100 in the run group. The exposure incidence was 12.8/1,000 hours of physical training in the walk group and 14.9/1,000 hours in the run group. There was no statistically significantly difference in the total number of injured recruits in the two groups (64 vs. 85, chi(2) = 2.90, p = 0.09, relative risk [RR] = 1.24). There were, however, significantly more lower-limb (43 vs. 75, chi(2) = 9.77, p = 0.0018, RR = 1.65) and knee injuries (15 vs. 35, chi(2) = 6.54, p = 0.011, RR = 2.14) in the Run group. Lower-limb injuries constituted 79.8% of all Run injuries and 61.1% of all Walk injuries. Injuries in the Run group produced more morbidity, with nearly double the number of days of restriction, hospitalization, and not fit for duty. Standardized morbidity rates showed an average of 5.4 days of restriction per injury in the Run group and 3.96 days of restriction per injury in the Walk group. Reduction of running distance in the physical training program resulted in significant reductions in both the incidence of lower-limb injury and the overall severity of injury.