This study was designed to test the hypothesis that continuation of a regular running and/or aerobics program during late pregnancy at or above 50% of preconceptional levels limits fetal growth. Accordingly, detailed neonatal morphometric data were gathered in the offspring of two groups: 77 well-conditioned recreational runners and aerobic dancers who were delivered at term after continuing their exercise regimen at or above 50% of the preconceptional level throughout pregnancy and 55 matched controls. Daily exercise performance was quantitated before conception and throughout pregnancy. Significant reductions in birth weight (-310 gm), birth weight percentile (-20), ponderal index (-0.24), its percentile (-30), and the fetoplacental weight ratio (-0.7) were seen in the offspring of the exercise group whereas crown-heel length (51.4 cm) and head circumference (35.0) were similar in the two groups. Reductions in two-site skin-fold thickness (-1.5 mm), skin-fold percentile (-30), calculated percent body fat (-5.0%), and fat mass (-220 gm) in the offspring of the exercise group confirmed the asymmetric pattern of growth restriction and indicated that approximately 70% of the difference in birth weight could be explained by the difference in neonatal fat mass. In runners, the relative level of exercise performance in the last 5 months of pregnancy explained 40% of the variability in birth weight over an 1100 gm birth weight range. We conclude that continuation of a regular aerobic or running program at or above a minimal training level during late pregnancy results in an asymmetric pattern of growth restriction that primarily impacts on neonatal fat mass.