Different methods, both experimental and numerical, to determine the flammability limits are compared and evaluated, exemplified by a determination of the flammability limits of methane/hydrogen/air mixtures for hydrogen fuel molar fractions of 0, 0.2, 0.4 and 0.6, at atmospheric pressure and ambient temperature. Two different experimental methods are used. The first method uses a glass tube with visual observation of the flame, whereas the second method uses a closed spherical vessel with a pressure rise criterion to determine whether flame propagation has occurred. In addition to these experiments, the flammability limits are determined numerically. Unsteady planar and spherically expanding flames are calculated with a one-dimensional flame code with the inclusion of radiation heat loss in the optically thin limit. Comparison of the experimental results with the results of the planar flame calculations shows large differences, especially for lean mixtures. These differences increase with increasing hydrogen content in the fuel. Better agreement with the experimental results is found for the spherically expanding flame calculations. A limiting burning velocity of 5 cm/s is found to predict the upper flammability limit determined with the tube method very well, whereas the limiting flame temperature approach was found to give poorer agreement. Further analysis indicates that the neglect of flame front instabilities is the probable cause of the large differences between experimental and numerical results at the lower flammability limit.