In order to estimate the influence of the fuel composition on speciated hydrocarbon emissions from gasoline engines a model has been developed for the processes undergone by the fuel which escapes the main combustion event. One of the most important ways that this occurs is by trapping in crevices followed by mixing and partial oxidation with the hot burnt gas during the power and exhaust strokes. This complex process has been modelled by recognising some important characteristics. It is observed that the fraction of a fuel species emitted is well correlated with its rate constant for reaction with OH radicals and that this is independent of the rest of the fuel composition. This means that (a) the chemistry is significant (not just mixing) and (b) the radicals carrying out the oxidation originate from the burnt gases. The decoupling of radical concentrations from the fuel composition considerably simplifies the modelling.Using this approximation a model for the hydrocarbon oxidation has been developed with detailed chemical kinetics but simple physical conditions. This correctly predicts the main partial oxidation products and the ratios of the emissions for a range of single component fuels (propane, butane, isooctane) in both test engines and cars.A start has been made in extending this model to incorporate the time dependent temperatures, pressures and release rates of hydrocarbons from crevices. The amount of radicals present in the burnt gases, and hence the extent of hydrocarbon burnup is found to depend strongly on the amount of combustion generated NOx, which catalyses the disappearance of the radicals in the burnt gas during the power stroke. Including oxidation under higher temperature conditions with a correspondingly larger mole fraction of radicals leads to a better prediction of ethyne and Co.