Epidemiologic studies of potential behavioral influences on preterm birth have proliferated and yet, with the exception of tobacco smoking, none can be considered an established cause. We conducted a comprehensive summary of the epidemiologic evidence on tobacco, alcohol, and illicit drug use, and physical, sexual, and occupational activity, to identify promising research directions, as well as research topics that are at an impasse based on currently available methods. Tobacco use is weakly but consistently associated with preterm birth-stronger for heavier smoking, and for spontaneous preterm birth and earlier preterm births. Weaker evidence suggests an adverse effect of environmental tobacco smoke, heavy alcohol or cocaine use, and physically strenuous work. Low levels of alcohol use, caffeine, sexual activity, and employment have generally not been found to be associated with preterm birth, and leisure-time physical activity has generated mixed results. Progress will require more detailed consideration of antecedents, new technologies for assessing exposure, and examination of biologic consequences of the behaviors of interest, focusing on pathways thought to mediate preterm delivery. New strategies-rather than more applications of the same approaches used in past studies-will move the research toward identifying causal relationships and, ultimately, may suggest preventive measures.