Five healthy fasting male subjects were each given single doses of intravenous ampicillin (471 mg), oral ampicillin tablets (495 mg), oral bacampicillin hydrochloride tablets (562 mg ampicillin equivalent), and oral pivampicillin hydrochloride capsules (491) mg ampicillin equivalent) in a crossover experiment. The resulting concentrations of ampicillin were determined in plasma and urine. The pharmacokinetic analysis was made according to a two-compartment open model. The total distribution volume of unbound ampicillin during the disposition phase was 0.247 +/- 0.045 (SD) liter/kg, which is only slightly more than the extracellular fluid, suggesting that tissue binding and intracellular distribution of ampicillin are limited. The bioavailability of the esters bacampicillin (86 +/- 11%) and pivampicillin (92 +/- 18%) was significantly greater than that of ampicillin (62 +/- 17%); however, the difference between the esters was not statistically significant. The adsorption for all drugs given orally proceeded at a constant rate, suggesting zero-order release rates from the products. The adsorption rate was highest for bacampicillin (0.89 +/- 0.39 of dose absorbed per minute), followed by pivampicillin (0.64 +/- 0.19) and ampicillin (0.58 +/- 0.16). Bacampicillin also had the shortest lag time for the start of absorption (7.0 +/- 0.9 min) under the present conditions. Thus, in comparison with ampicillin, the esters have a higher bioavailability, which, in fact, is close to the theoretically highest possible value by clearance concepts. The higher bioavailability in connection with higher absorption rates may be clinically important in ampicillin treatment by the oral route.