Studies of plasma concentrations and the application of pharmacokinetic parameters in human smoking published during recent years are reviewed. It is clear that this approach contributes to better understanding and to the quantitative evaluation of many problems related to human smoking, such as: first, smoking behavior, second, self-regulation and compensation when switching between different brands and third, relationships between in vitro machine smoking yields and actual human uptake. The reviewed articles show that the widely believed opinion that cigarettes with low yields, when in vitro tested, are less hazardous, must be questioned. For the majority of cigarette brands there is no defensible correlation between in vitro analytical machine yields and actual uptake and bioavailability. Based on these findings suggestions are given to modify the use of tar and nicotine yields obtained by in vitro machine testing methods for classification purpose.