The purpose of present study was to assess the relationship between anaerobic threshold (AT) and performances in three different distance races (i.e., 5 km, 10 km, and 10 mile). AT, VO2 max, and related parameters for 17 young endurance runners aged 16--18 years tested on a treadmill with a discontinuous method. The determination of AT was based upon both gas exchange and blood lactate methods. Performances in the distance races were measured within nearly the same month as the time of experiment. Mean AT-VO2 was 51.0 ml . kg-1 . min-1 (2.837 l . min-1), while VO2 max averaged 64.1 ml . kg-1 . min-1 (3.568 l . min-1). AT-HR and %AT (AT-VO2/VO2 max) were 174.7 beats . min-1 and 79.6%, respectively. The correlations between VO2 max (ml . kg-1 . min-1) and performances in the three distance races were not high (r = -0.645, r = -0.674, r = -0.574), while those between AT-VO2 and performances was r = -0.945, r = -0.839, and r = -0.835, respectively. The latter results indicate that AT-VO2 alone would account for 83.9%, 70.4%, and 69.7% of the variance in the 5 km, 10 km, and 10 mile performances, respectively. Since r = -0.945 (5 km versus AT-VO2) is significantly different from r = -0.645 (5 km versus VO2 max), the 5 km performance appears to be more related to AT-VO2 than VO2 max. It is concluded that individual variance in the middle and long distance races (particularly the 5 km race) is better accounted for by the variance in AT-VO2 expressed as milliliters of oxygen per kilogram of body weight than by differences in VO2 max.