Many studies have estimated nicotine intake during smoking from the analysis of metabolites of nicotine in body fluids, particularly cotinine in saliva and nicotine and up to eight metabolites in urine. In addition, estimates of the mouth intake of nicotine during smoking have been made using the analysis of smoked filters. Each of these approaches has merits, but to date, very few publications have compared these methods of estimating smoke intake. Consequently, an in-patient study was conducted with 75 smokers of 1-17 mg FTC tar products smoking their own brands. Smokers of cigarettes from four tar bands were recruited to assure a wide range of intake: 1-3 mg (15 subjects); 4-6 mg; 7-13 mg; 14+ mg (20 subjects each). The subjects stayed in a clinic for 5 days during which time they were allowed to smoke in a smoking room whenever they wished. The filter from each cigarette had to be returned to a clinician to enable the issue of a subsequent cigarette. These filters were then analyzed to estimate the daily mouth intake of nicotine. Five 24-hour urine samples per subject were collected and analyzed for nicotine, cotinine, 3-OH cotinine and their respective glucuronide conjugates. Daily saliva samples were collected at 18.30 each day for cotinine analyses. On the fourth day, additional saliva samples were collected 08.30 and at 13.30 to assess any diurnal variations in saliva cotinine levels. Each of the methods correlated significantly (p<0.01) with the other two, but the correlation between nicotine mouth intake and urinary output was far better than the saliva cotinine correlations. Averaging the results over 5 days improved the mouth intake/urinary output correlation even further but had little effect on the saliva cotinine correlations. Multiple regression analysis implies that urinary output is an amalgam of the nicotine input from multiple days.