Tobramycin dosage regimens were individually calculated from serum concentration-time data in 64 patients. Tobramycin pharmacokinetic parameters and dosage requirements demonstrated wide interpatient variation. The tobramycin half-life varied from 0.5 to 8.6 hours in patients with normal serum creatinines. The mean (+/- S.D.) distribution volume was 0.22 (+/- 0.09) liter/kg for all patients. Dosage requirements were higher for the younger patients, however, considerable variability existed within age groups. In patients with normal serum creatinines, an 8-hour dosing interval was optimal in only 17 of the 46 patients. A multiple regression model using age, weight, and creatinine clearance could explain only 44.2% of the variance in tobramycin clearance. Measured steady-state peak and trough concentrations compared favorably with desired peak and trough concentrations. The mean (+/- S.D.) desired peak was 7.2 (+/- 1.8) microgram/ml, and the mean (+/- S.D.) measured peak was 7.1 (+/- 2.0) microgram/ml. The mean (+/- S.D.) desired trough was 1.1 (+/- 0.9) microgram/ml, and the mean (+/- S.D.) measured trough was 1.3 (+/- 0.8) microgram/ml. Many patients required dosage regimens exceeding those commonly recommended; however, no cases of ototoxicity or nephrotoxicity were encountered. This model proved successful in calculating dosage regimens of tobramycin to obtain optimal serum concentrations.