It has been said that people smoke cigarettes for nicotine but die from the tar, and it is generally acknowledged that the cigarette is an extraordinarily dirty drug delivery system. However, despite the fact that nicotine itself is not a particularly toxic drug in smoking doses, harm reduction approaches have had relatively little emphasis in tobacco control. The strategy of gradually and progressively reducing machine-smoked deliveries of tar and nicotine has been demonstrated to be entirely bankrupt, due to the ease, intended by tobacco manufacturers, with which smokers can achieve much higher intakes from lowtar cigarettes than the nominal yields would suggest. To have any chance of succeeding, policies for cigarette product modifIcation must be fIrmly grounded in an understanding of the .factors driving smoking behaviour. This presentation will outline evidence on the role of nicotine in tobacco smoking and the arguments for and against the two main contending strategies: reduce the absolute-bioavailability of nicotine from cigarettes to the point where they will no longer be reinforcing; or leave nicotine largely unregulated while reducing emissions of gas phase and tar components.