A 5-year longitudinal cohort study of the running development of Japanese adolescents was conducted at a secondary school in Tokyo from 1968 to 1994. The purpose was to investigate changes in the pattern of development over time. The test items were an endurance run (1,500 m for boys and 1,000 m for girls) and a dash (50 m for both sexes). The total number of subjects was 512 boys and 516 girls in the endurance run, and 825 boys and 895 girls in the dash. These students entered the first grade of the school from 1968 to 1989 at the age of 12 years, and were followed for the next 5 years. In the analyses, a quadratic regression model was used to characterize intra-individual, age-related changes in running performance. The model fit the observed data, and the validity of the model was confirmed. Secular changes in the running development of boys were minimal over the last 30 years. Changes in girls, however, had three different patterns: 1) an overall decrease in performance with an almost unchanged estimated performance in the first grade, 2) a greater decrease in the late-teenage years, and 3) a decrease in the age of maximal performance. Secular changes in physical growth (stature and body mass index) could not fully explain the secular changes in the running development of girls, suggesting that the underlying causes may be social and cultural. Am. J. Hum. Biol. 10:765-779, 1998. Β© 1998 Wiley-Liss, Inc.