The smoke intake of 865 undisturbed smokers of over 10 cigarettes per day was measured using plasma nicotine and cotinine, and expired carbon monoxide (CO) as markers. While nicotine yields, according to Federal Trade Commission (FTC) analytical standards, varied 16-fold from 0.1 to 1.6 mg/cigarette, the corresponding plasma nicotine values varied from around 25 to 45 ng/ml, and estimated mean nicotine intake of smokers varied from around 0.75 to 1.25 mg/cigarette. Expired CO and plasma cotinine values also varied in similar proportion, but mean daily cigarette consumption was independent of the FTC nicotine yield of the cigarettes smoked. The results indicate that pharmacodynamic satiation causes behavioral regulation, and that smokers of very high yield brands compensate downward, and vice versa. The ratio of tar yield to nicotine yield usually increases with increasing tar yield; therefore tar intake is likely to increase at higher tar yields, even though the increment of nicotine intake is small. It follows that FTC analytical determinations are poor predictors of relative intake of nicotine, CO, or tar, while rankings based on mean tar-to-nicotine ratio of a brand's smoke could be more meaningful. Moreover, the considerable variation of individual smoking behavior suggests that precise numerical rankings of cigarettes are not justified. An analogic ranking of cigarettes into a few broad classes would better reflect the realities and expectations of average consumers.