Two versions of a detailed chemical kinetic model of n-butane autoignition are proposed. The key distinctions between the versions are the exclusion (Model I) or inclusion (Model II) of the direct abstraction or apparently bimolecular path of the reaction HO 2 + HO 2 → H 2O 2 + O 2, and the selection of rate parameters for the reaction C 2H 5 + O 2 → C 2H 4 + HO 2. Both models (versions) were evaluated over 1200–1400 K for a stoichiometric mixture and over 720–830 K for lean to rich mixtures by comparing the computed autoignition delays with those of two experiments: (1) a shock-tube experiment (representative of high-temperature chemistry) and (2) a rapid compression experiment (representative of low-temperature chemistry). The experiments demonstrate the ablity of the models to predict autoignition delays at high pressures typical of automobile engines. Four interpretations of the rapid compression data using the present model are also described. The two models provide the same behavior of autoignition delay (a macroscopic phenomenon), but the sensitivities of the delays to reaction rate constants (a microscopic aspect of the autoignition mechanism) are remarkably different. Therefore, while the two models are shown to reproduce the macroscopic experimental data, more research is needed to determine which model is valid at the microscopic level.