There is a strong, significant relationship between coffee consumption and smoking. In six epidemiological studies reviewed and analyzed here, 86.4% of smokers consumed coffee versus 77.2% of nonsmokers. Exsmokers use more coffee than nonsmokers but somewhat less than smokers. Seventeen experimental studies suggest that the pharmacologic effect of caffeine in coffee may be partially but not totally responsible for the relationship. Conditioning, a reciprocal interaction (caffeine intake increases anxiety/arousal--nicotine decreases it), or joint effect of a third variable (e.g., stress, alcohol) may account for the relationship. In abstinent smokers, blood caffeine levels increase and remain elevated for as long as 6 months. These higher caffeine plasma levels may be sufficient to produce caffeine toxicity syndrome. A review of 86 studies of nicotine withdrawal, caffeine withdrawal, and caffeine toxicity suggests that the symptoms are similar enough to be confused, and that reported nicotine withdrawal symptoms may be a mixture of nicotine withdrawal and caffeine toxicity.