The effect on subsequent pregnancy outcome related to abortion was studied in a sample of 26,000 births in Taiwan. When no adjustment was made for confounding maternal variables, significant relationships were found between previous abortion and various problems of outcome of subsequent pregnancies. However, in a matched-pair cohort study controlling for relevant maternal characteristics, all such relationships disappeared. It is concluded that the deleterious effects of abortion on subsequent pregnancy outcome reported by others may be a consequence of the failure to adjust for maternal factors which are related both to abortion status and pregnancy outcome, and not to abortion itself.