The feasibility of collecting DNA through the mail from a cohort of current and former smokers was assessed. Also examined was whether monetary incentives would increase response rates. A random sample of 300 subjects, stratified by 20 U.S. communities, was selected to participate. The sampling frame included the 6,726 people who were in both the Community Intervention Trial for Smoking Cessation (COMMIT) between 1988 and 1993 and the follow-up study in 2001, and who consented to being contacted again. Subjects were further randomized within communities to incentive arms of 10 US dollars, 2 US dollars, or 0 US dollars. A total of 110 usable samples were returned (37%), and the 10 US dollars incentive arm had the highest response (43%). Logistic regression revealed no significant predictors of sending a DNA sample, although in a larger study, similar-sized odds ratios would be statistically significant for subjects who received the 10 US dollars incentive and for those who were White, female, or college graduates or whose household incomes were more than 60,000 US dollars per year. The spectrophotometer-determined median DNA yield was 44.93 microg (range=4.00-425.86 microg). Assuming that 50 ng of DNA would be needed for polymerase chain reaction amplification to determine any given genotype, 80-8,517 runs would be attainable. Qualitative findings suggest several methodological improvements to boost response rates. Institutional review board requirements, which are standardized on the inpatient, clinical protocol model, stipulated that noninstitutionally based subjects needed a witness to initial and date every page as well as sign the consent form. This pilot study showed that this requirement could pose some challenges in population-based research.