A recently proposed bioavailability estimation procedure, the "semisimultaneous" method, in which the test and reference dose administrations are separated by a short time interval and total concentrations are analyzed, was compared with the stable isotope method for precision and accuracy. By administering isotope-labeled (reference) and unlabeled (test) terbutaline in a semisimultaneous fashion, the bioavailability could be determined with both methods at the same time. The extent and rate of bioavailability of oral terbutaline was determined in eight healthy volunteers by use of both model fitting, AUC methods, and deconvolution. The AUC ratio and the deconvolution methods, by use of the separate isotope data, gave bioavailability estimates of 14.5% +/- 4.1% and 12.2% +/- 3.9%, respectively. According to the semisimultaneous method, bioavailability estimates with the same data sets were 11.8% +/- 4.5% obtained from fitting a model to the data and 11.0% +/- 3.7% by use of a combined model fitting-deconvolution procedure. An excellent agreement between the semisimultaneous and the stable isotope methods was also obtained in the estimation of the rate of absorption.