Placental and hepatic xenobiotic-metabolising activities were studied in smokers and non-smokers, who were classified by anamnestic interview, plasma thiocyanate and plasma cotinine determinations. Plasma thiocyanate assay is inadequate in the classification of smokers and non-smokers. Plasma cotinine levels reflect more accurately the smoking status. The anamnestic smokers remained smokers and several self-declared non-smokers proved to be smokers. On the basis of plasma cotinine determination all real smokers had higher 7-ethoxyresorufin 0-deethylase (ERDE) activities measured either in placental microsomes or liver biopsy homogenates than non-smokers. Classification based on plasma cotinine levels showed a statistically significant (P less than 0.001) difference between smokers and non-smokers in liver homogenate ERDE activity. However, cotinine levels did not correlate with any of the xenobiotic-metabolising activities tested. An objective biochemical marker, such as cotinine determination seems to be necessary when evaluating the effect of smoking on drug metabolism in man.