Despite widespread belief that children are aerobically trainable, studies examining the ability of a child to improve maximal oxygen uptake (VO2max) have yielded inconsistent findings. The present investigation, using meta-analysis, examined the effects of physical activity, gender, experimental design, and sufficiency of exercise on the VO2 max of child subjects. Sixty-nine studies examining the effects of training on children were originally located; 28 met criteria for inclusion. From these studies, 70 effect sizes (ESs) were calculated. Some studies employed cross-sectional (XS) designs involving comparisons of intact groups of subjects; the others used a pretest-posttest (PP) design, which followed subjects throughout a specified training program. Average ES indicated a considerable difference between trained and untrained subjects though several possible sources of confounding (e.g., subject self-selection) in XS studies were identified. Effect sizes of .94 (+/- 1.00) and .35 (+/- 0.82) were achieved for XS and PP designs, respectively. Further analyses were conducted with the PP design studies. In these studies, subjects improved approximately 2 ml.kg 1 x min-1. In the PP studies, effect sizes were not significantly affected by (a) gender, (b) "sufficient" and "insufficient" training protocol, or (c) test mode. Results indicated that reported changes in VO2max in children are small to moderate and are a function of the experimental design used.