To estimate daily nicotine consumption from the smoking of cigarettes delivering different amounts of nicotine, we studied 12 healthy subjects who smoked 30 cigarettes a day of their usual brand (x = 1.2 mg nicotine) or high- (2.5 mg) or low-nicotine (0.4 mg) research cigarettes. Blood nicotine and carboxyhemoglobin concentrations were measured every 2 hr. Nicotine consumption was estimated by the 24-hr area under the blood concentration-time curve (AUC) and compared across smoking conditions. There was considerable interindividual variation in the nicotine AUC among subjects smoking research cigarettes or while smoking usual brands, even when the latter were normalized on the basis of machine-predicted nicotine delivery. Most subjects smoked the high-nicotine research cigarettes less intensively so that nicotine levels were only modestly higher after smoking high-nicotine cigarettes than after usual brands. Low-nicotine research cigarettes were not smoked more intensively than usual brands and blood nicotine levels were substantially lower than those after smoking a usual brand. Nicotine consumption while smoking usual brands correlated strongly with consumption while smoking high- (r = 0.91) and low-nicotine (r = 0.85) research cigarettes. Circadian studies of blood concentration of nicotine while smoking cigarettes provided a direct estimate of the level of nicotine in the body throughout the day. Results confirm observations by others that levels of nicotine in the body vary widely among individuals even when smoking the same number of identical cigarettes. Thus, neither number of cigarettes smoked nor smoking-machine delivery predict daily nicotine exposure very well.