In this study, using meta-analysis the findings of all individual reports on the effects of exercise training on pregnancy outcomes were pooled to determine the effects of such training on the pregnant woman and her fetus. Meta-analysis is a statistical technique used to summarize the results of experimental studies that address a common problem. Results show that women who exercised during their pregnancies did not differ from sedentary women for any of the outcome variables measured: maternal weight gain (P = 0.07), infant birth weight (P = 0.20), length of gestation (P = 0.67), length of labor (P = 0.14), and APGAR scores (P = 0.59). Many of the exercise programs exceeded, without apparent adverse effects, the recommended limitations set by the American College of Obstetricians and Gynecologists. Overall, an exercise program using any of a variety of exercise modes that is performed for an average of 43 min.d-1, 3 times.wk-1, at a heart rate of up to 144 bpm, does not appear to be associated with adverse effects to the mother or fetus in a healthy normal pregnancy. However, these findings should be cautiously applied owing to the nature of the currently available data base. Recommendations or precautions for programs of greater intensities can not be made at this time.