This study used a short-term laboratory model of smoking cessation and relapse to prospectively examine the effects of programmed self-administered smoking re-exposure during early abstinence. Sixty-seven subjects who had quit smoking for 3 days were randomly assigned either to smoke five cigarettes in their natural environment or to remain abstinent during the exposure period. The main hypothesis, that relapse to regular smoking would be quicker and more prevalent in exposed subjects, was supported. All exposed subjects had relapsed by 2 days post-exposure while 16% of unexposed subjects remained continuously abstinent throughout the 8 day study. This behavioral effect was seen in spite of acute decreases in reported desire to smoke and increases in guilt measured just after exposure. The study supports a role for stimulus re-exposure effects in the relapse process and suggests that additional research on experimental re-exposure is warranted.