To optimize the efficiency and safety of mixtures of nonozone-depleting refrigerant replacements, precise measurements of the lean flammability limits of pure refrigerants and the critical flammability ratio of their mixtures are crucial. Current test methods that model the accidental ignition of a volume of premixed fuel and air provide ambiguous results when measuring the limiting behavior of weakly flammable refrigerants. An alternative approach using the extrapolation of the global extinction stretch rates to zero as measured in a premixed counterflow burner has been tested. In this work the approach, accuracy, and precision of the technique as it applies to CH2F2 (difluoromethane, R-32) are presented. Comparisons are made to the behavior of CH4 in the same burner and to published data on the lean limiting equivalence ratio, Phi(0). The lean flammability limit of CH2F2 in dry air is found to be Phi(0) = 0.78 +/- 0.04. Concurrent computational modeling of the combustion of a one-dimensional, unstrained flame of CH2F2 in air, individually and in mixtures with CH4, has been performed. Initial estimates of the laminar flame speed for CH2F2/air mixtures from stoichiometric to lean are reported and interpreted in light of the experimental results.