The rate of nicotine transfer to a smoker and the site of absorption within the human respiratory tract may depend on the distribution of nicotine between the vapour and particulate phases of mainstream smoke. A technique has been developed to measure the evaporation of nicotine from mainstream smoke using a denuder tube. The denuder tube, a long narrow glass tube, is designed to ensure laminar flow and minimise particle diffusion or deposition. This device selectively collects vapour phase nicotine while allowing the particulate phase nicotine to pass through. Nicotine vapour reaching the walls reacts with a coating of oxalic acid and is subsequently analysed by gas chromatography. Nicotine which has remained on the particles is collected by a Cambridge pad at the exit of the denuder tube. Initial experiments with nicotine vapour in the absence of particles confirmed literature values for the diffusion coefficent of nicotine. However, experiments with mainstream smoke found that less than 1% of the nicotine was initially in the vapour phase and an apparent diffusion coefficient in smoke about 40 times lower than that for nicotine vapour alone. It was also found that the proportion of nicotine evaporating and diffusing from mainstream smoke increased as the smoke was diluted. Tobacco blend type and ‘tar’ to nicotine ratio were also found to influence nicotine evaporation.