This study will test whether abstinent smokers who experience intense withdrawal are more likely to relapse to smoking than abstinent smokers who experience little withdrawal. Most smokers who quit on their own relapse in the first two weeks. Whether this is due to aversive withdrawal symptoms (e.g. irritability) is unknown because prior studies have had several major flaws; e.g., retrospective designs, nongeneralizable samples and unvalidated, self-report measures of withdrawal. We have experimentally documented several physiological signs and observable behaviors that reliably occur upon cessation of smoking and have the characteristics of classical withdrawal syndromes. We now propose to prospectively test the ability of these withdrawal variables to predict relapse in self-quitters. 1000 smokers who are about to quit on their own will name a reliable observer to rate their withdrawal behavior. These observer ratings plus weight, heart rate, and self-reported symptoms will be recorded prior to and 2, 14 and 30 days after subject quit. One year later, salivary cotinine will verify whether smokers have relapsed. We will test the hypothesis that among smokers who quit for 2 days, those with more intense withdrawal are more likely to relapse by one year follow-up. The results of this study will be a crucial, direct test of the nicotine dependence theory of smoking. The results will also encourage or discourage the development of smoking treatments designed to combat withdrawal symptoms.