Seventy subjects smoking their usual brand of cigarettes participated in a study to examine nicotine uptake as an index of the uptake of other smoke constituents and to assess the short-term intra-individual and inter-individual variability in smoke exposure (uptake). After orientation, subjects participated in a total of four test sessions (two consecutive days in the first week and two consecutive days two weeks later) and collected 24-h urine and buccal cells (mouthwash) at each session. Urine was analyzed for biomarkers of exposure that included metabolites of nicotine (free and total nicotine, cotinine, and 3'-hydroxycotinine), NNK (free and total NNAL), benzene (S-phenylmercapturic acid [SPMA]), acrolein (3-hydroxypropylmercapturic acid [HPMA]), butadiene (1-hydroxymethyl-2-propenylmercapturic acid [MHBMA], 3,4-dihydroxybutylmercapturic acid [DHBMA]), and pyrene (total 1-hydroxypyrene [1-OHP]). Buccal cells were analyzed for biomarkers of biologically effective dose of carbonyls (carbonyl protein adducts) and oxidative stress (8-hydroxy-2'-deoxyguanosine [8-OHdG]). Urinary biomarker data were expressed as urine concentration, output per day, and normalized to creatinine. Buccal cell biomarker data for carbonyl adducts were normalized to total protein and, for 8-OHdG, were expressed as concentrations and normalized to tryptophan. Data were log-transformed prior to statistical analysis to better accommodate extreme observations and skewness generally found in the data. Total nicotine uptake (molar sum of nicotine and metabolites) was strongly correlated (r > 0.7) with NNK and acrolein metabolites, was moderately correlated (r = 0.3-0.5) with metabolites of benzene, pyrene, and butadiene, and was not significantly correlated with carbonyl protein adducts or 8-OHdG. Contribution to total variation within a week (i.e. , day-to-day) was generally more than between weeks. With few exceptions, subject-to-subject variability was by far the greatest source of variation for all biomarkers. The contribution to total variance from between-week variation approached or exceeded the contribution for between-subject variation in only two cases, for MHBMA and 1-OHP in ultra-low "tar" smokers.