1 Extent and rate of absorption of erythromycin were studied in 24 healthy volunteers whose disposition kinetics after i.v. injections had been previously documented. 2 Two clinically attractive oral dosage regimens were administered: erythromycin stearate tablets 1 h before meals (Regimen A), and erythromycin base capsules 30 min after start of meals (Regimen B), each equivalent to erythromycin 250 mg, 6 h apart for 9 doses. 3 Serum concentrations of erythromycin measured during the 1st and 9th (steady-state) dosing intervals resulted in higher maximum serum concentrations for Regimen A (median 1.1, range 0-3.3 and 2.7, 0.6-7.3 mg/l for Doses 1 and 9, respectively) compared with Regimen B (0.4, 0-2.2 and 1.4, 0.2-4.9 mg/l). 4 Absorption occurred earlier with Regimen A with times to maximum concentrations (median, range) being 128, 60-greater than 360 and 118, 75-210 min for doses 1 and 9 respectively, (lag times 75, 15- greater than 360 and 73, 10-110 min) compared with 303, 130-greater than 360 and 173, 45-greater than 360 min (lag times 183, 70-greater than 360 and 190, 20-330 min) for Regimen B. 5 Where it could be assessed, absolute bioavailability for Regimen A was approximately 30% (Dose 1) and 65% (Dose 9) and 40% for both doses of Regimen B. 6 Whereas individual serum concentration-time curves were accurately predicted by the mean for Regimen A, predictability for Regimen B was impossible due to prolonged and variable lag time. 7 The large intersubject variability in erythromycin serum concentration after oral administration, has been shown conclusively to be related to variability in absorption kinetics and absolute bioavailability rather than to variability in disposition kinetics.