Ninety-two women in preterm labor were randomly assigned to either terbutaline or ethanol. In women with intact membranes, terbutaline was significantly more effective than ethanol in preventing progressive cervical dilation during the first 36 hours of treatment while betamethasone was given to the mother. Additionally, pregnancy was maintained significantly longer in women with intact membranes who were treated with terbutaline (15 +/- 4 days) rather than with ethanol (10 +/- 3 days). However, only 18% of women in each treatment group maintained their pregnancy beyond 36 weeks' gestation. In women with ruptured membranes and cervical dilation less than 4 cm, terbutaline was significantly better than ethanol in maintaining pregnancy for a minimum of 36 hours. Serious maternal side effects were not observed with terbutaline or ethanol, although a majority of women also received betamethasone. Neither drug caused serious adverse neonatal effects.