We evaluated a multicompartment analysis of forced expiration, based on modelling the lung as a set of twenty parallel compartments emptying exponentially with time constants ranging from 0.1-10 s; the forced expired volume signal was represented by a histogram showing the fraction of forced vital capacity as a function of compartmental time constants. We applied this technique to 80 healthy and 12 asthmatic subjects. The histograms computed from three consecutive forced expirations were poorly reproducible in 18 of the 80 healthy and 2 of the 12 asthmatic subjects. In the asthmatics, the time constant histograms conveyed no additional information on bronchial obstruction, beyond that already present in standard spirometric indices. A simulation study showed a high sensitivity of the histograms to the truncation of the terminal part of forced expiration. We conclude that the usefulness of the time constant histogram technique appears doubtful.