Studies have demonstrated the effectiveness of smoking cessation interventions targeted at women who are smoking during pregnancy. In contrast, there is limited information about the experience of smokers who have stopped before entering prenatal care--"spontaneous quitters." These women constitute the majority of women who stop smoking sometime during pregnancy, although evidence suggests that as many as one third relapse prior to delivery. We report the results of a population-based randomized clinical trial that tested the effectiveness of a relapse prevention program for spontaneous quitters. The intervention consisted predominantly of printed materials received through the mail. The population (n = 171) of spontaneous quitters was an ethnically diverse group of women enrolled in a large health maintenance organization. Biochemical confirmation of continuous abstinence through delivery revealed that 16% of the women in the experimental self-help program relapsed compared with 20% of usual care controls (NS). Analysis confirmed that the program was equally ineffective among all subgroups including women at highest risk for relapse. Given the negative outcomes associated with self-help materials and clinic-based counseling reported in this and other trials, alternative intervention strategies need to be developed and tested for this significant group of prepregnancy smokers.