The purpose of this study was threefold: to determine (a) the test-retest reliability of the 20-m shuttle test (20 MST) (number of laps), (b) the concurrent validity of the 20 MST (number of laps), and (c) the validity of the prediction equation for VO2max developed by Léger, Mercier, Gadoury, and Lambert (1988) on Canadian children for use with American children 12-15 years old. An intraclass coefficient of .93 was obtained on 20 students (12 males; R = .91 and 8 females; R = .87) who completed the test twice, 1 week apart (MT1 = 47.80 +/- 20.29 vs. MT2 = 50.55 +/- 22.39 laps; p > or = .13). VO2peak was obtained by a treadmill test to volitional fatigue on 48 subjects. The number of laps run correlated significantly with VO2peak in males (n = 22; r = .65; F [1, 20] = 14.30 p < or = .001), females (n = 26; r = .51; F [1, 24] = 8.34; p < or = .01), and males and females = (r = .69; F [1, 46] = 42.54, p < or = .001). When the measured VO2peak (M = 49.97 +/- 7.59 ml.kg-1.min-1) was compared with the estimated VO2max (M = 48.72 +/- 5.72 ml.kg-1.min-1) predicted from age and maximal speed of the 20 MST (Léger et al., 1988) no significant difference was found, t (47) = -1.631; p > or = .11, between the means; the r was .72 and SEE was 5.26 ml.kg-1.min-1.(ABSTRACT TRUNCATED AT 250 WORDS)