The Armed Forces considered introducing compulsary physical training for all officeres, and replacing performance tests by that training for some categories. To investigate how that model would work, five units (ca 700 officers) were selected as experimenatal units (E) while five other units were selected as controls (C) (ca 500 officers). For medical examination and testing of physical performance all female officers and 25% of the male officers were selected. All were selected for a questionaire regarding physical activity habits. The implemantation was sucessful in some units but not in others. The majority of the officers were in favour of including the training in their weekly planning. Both E and the C reported increased physical activity (E more than C) during the course of the study. E did not differ from C regarding aerobic capacity, muscle strength, BMI, cholesterol, hip/waist circumference ratio. Comparing officers with a sample of the Swedish population, showed that among officers, overweight was equally prevalent in males, while female officers showed a lower prevalence; high blood pressure as well as too low maximal oxygen uptake values, were less frequent among officers; physically active was higher. However, 1 of 5 officers answered that they trained less than once a week. The absence of performance tests did not have a negative influence on training volume, endurance, and muscle strength. The majority of the officers had a positive opinion of the new model. However, to sucessfully implement the new model, considerable educational efforts are needed.