Two nonlinear pharmacokinetic models were simulated to investigate the relationship between single and multiple dose bioequivalency parameters for drugs such as phenytoin and propranolol which exhibit either saturable elimination kinetics or a capacity limited first pass effect. Mean Tmax, Cmax and area under the plasma-concentration time curve values from 0 to infinity (AUC 0-infinity) were compared after a single and multiple dose(s) of a test or reference drug. The aim was to determine if there were systematic changes in the limits of the single dose confidence interval at steady state that would limit the usefulness of confidence intervals following a single dose in accurately predicting bioavailability following multiple dosing. The 90 per cent confidence interval expressed as a percentage of the reference mean for Tmax, Cmax, and AUC 0-infinity showed model dependent changes from single to multiple dosing in response to the level of data error and changes in absorption. Changes in clearance also seemed to have a marked effect on the observed limits of the single and multiple dose confidence intervals especially for Cmax which showed a characteristic change in the intervals as a function of the clearance ratio. The model used to describe phenytoin had confidence intervals for Cmax and AUC 0-infinity from single to multiple dosing that were similar to that seen for the experimental data. However, the model predictions for Tmax confidence intervals following single and multiple dosing was at variance with the experimental data for formulations A and B.