Forced expiratory manoeuvres are extensively recorded using flow meters coupled to pressure transducers, which usually behave like second-order filters. To assess what should be the dynamic characteristics of such equipment for accurate determination of common forced expiration indices, 125 curves were obtained from 25 healthy subjects in the best technical conditions. The flow-time curves were then submitted to various degrees of second-order filtering, and the indices derived from filtered and unfiltered curves were compared. Considering that experimental error is acceptable if it does not exceed 20% of the normal interindividual variability for 95% of the curves, the following conclusion was drawn; with an optimal damping ratio (r) of 0.7, forced expiratory volume in one second, maximum midexpiratory flow rate and maximal expiratory flow at 25% of the forced vital capacity (MEF25) are still correctly measured when the resonant frequency (fn) is as low as 3 Hz. The corresponding figures are 5 Hz for MEF50, 20 Hz for peak expiratory flow rate and above 20 Hz for MEF75. Higher values of fn are usually required when r is higher or lower than 0.7.