To test whether cigarettes with low tar, low carbon monoxide, and medium nicotine yield produce less dangerous effects than cigarettes low in tar and CO but high in nicotine, 12 subjects were recruited to smoke nicotine-enriched cigarettes. The subjects smoked three types of cigarettes in the three experimental conditions: (1) their own brand; (2) cigarettes with 4.8 mg tar, 4.0 mg CO, and 0.5 mg nicotine; (3) cigarettes with 5.8 mg tar, 4.1 mg CO, and 1.1 mg nicotine. Subjects monitored their daily consumption for 12 weeks; 4 weeks for each condition. During laboratory visits, the subjects smoked a cigarette while their heart rate and carbon monoxide in expired air were measured pre- and post-smoking. A blood sample was drawn and analyzed for nicotine and cotinine in each experimental condition. No significant differences in daily cigarette consumption were found, although a trend (P less than 0.07) in the direction of fewer nicotine-enriched cigarettes per day was found. Levels of CO varied significantly among the three conditions: The subjects' own brands yielded the highest level, while the nicotine-enriched cigarette yielded the lowest level. No differences were found for nicotine or cotinine levels. A second purpose of the experiment was to record the degree of nicotine titration displayed by individual smokers, tar and CO levels remained constant in the experimental cigarettes. No general titration effect was observed, although for daily consumption it approached significance. When the subjects' nicotine dependence, measured with a tolerance questionnaire, was taken into account, a correlation with daily consumption was found (r = 77, P less than 0.005). A cigarette with low tar and CO, but medium to high nicotine yield, would seem to produce less hazardous effects and is worthy of further investigation. The controversial question of whether smokers titrate for nicotine is a function of the individual's nicotine dependence.