One hundred women in preterm labor were randomly treated with ritodrine or terbutaline in a double-blind fashion. The drugs were comparably effective during intravenous therapy but, in women with intact membranes, an oral dose of terbutaline, 30 mg daily, was significantly more effective than ritodrine, 120 mg daily, in preventing recurrent labor during a 5-day course of oral therapy (one of 19 versus 12 of 23, p less than 0.001). In women with intact membranes, pregnancy was prolonged 40 +/- 25 days (mean +/- SD) in women receiving terbutaline orally and only 22 +/- 24 days in women receiving ritodrine orally (p less than 0.01). In women with intact membranes, a heart rate greater than or equal to 130 bpm occurred in in a higher proportion of women receiving intravenous treatment with ritodrine than among those receiving terbutaline (20 of 31 versus 8 of 27, p less than 0.05). Terbutaline-treated women, however, were significantly more likely to have a serum glucose level in excess of 140 mg/dl than were women treated with ritodrine (13 of 26 versus 6 of 29, p less than 0.05). Side effects commonly observed during intravenous therapy included nausea (22%), chest pain (15%), and shortness of breath (15%). Side effects were significantly (p less than 0.025) more likely to occur during periods when the infusion rate was being increased rather than during periods when the infusion rate was constant.