This paper reviews some existing models for calculating air:fuel ratio based on analysis of exhaust-gas emissions, as an alternative to the conventional direct measurement of the Row rates of air and fuel. A comprehensive model for accurate calculation is proposed, and a study is made of the model＇s sensitivity due to instrument errors in measuring emission concentrations. Results show that only four emission parameters-hydrocarbon, carbon monoxide, carbon dioxide and oxygen-are required for accurate calculation of air:fuel ratio. The sensitivity parameters of these emissions are low relative to the sensitivity parameters due to errors in measuring rates of fuel- and air-flow. The total absolute sensitivity of the former is about a tenth of the latter at stoichiometric air:fuel ratio. This means that to obtain an accurate air:fuel ratio for an IC engine, instruments used for measuring fuel-and air-flow rate should have at least ten times the accuracy of the emission analysers, especially for gasoline engines where equivalence ratio is determined at unity. A point is mooted: should a method based on exhaust emissions be regarded as the ＇standard method＇ for measuring the air:fuel ratio in gasoline engines?