To describe the relapse curve and rate of long-term prolonged abstinence among smokers who try to quit without treatment. Systematic literature review. Cochrane Reviews, Dissertation Abstracts, Excerpt Medica, Medline, Psych Abstracts and US Center for Disease Control databases plus bibliographies of articles and requests of scientists. Prospective studies of self-quitters or studies that included a no-treatment control group. Two reviewers independently extracted data in a non-blind manner. The number of studies was too small and the data too heterogeneous for meta-analysis or other statistical techniques. There is a paucity of studies reporting relapse curves of self-quitters. The existing eight relapse curves from two studies of self-quitters and five no-treatment control groups indicate most relapse occurs in the first 8 days. These relapse curves were heterogeneous even when the final outcome was made similar. In terms of prolonged abstinence rates, a prior summary of 10 self-quitting studies, two other studies of self-quitters and three no-treatment control groups indicate 3-5% of self-quitters achieve prolonged abstinence for 6-12 month after a given quit attempt. More reports of relapse curves of self-quitters are needed. Smoking cessation interventions should focus on the first week of abstinence. Interventions that produce abstinence rates of 5-10% may be effective. Cessation studies should report relapse curves.