Soft tissue infection models have been used to study both the postantibiotic effect (PAE) and the effect of dosage intervals on antimicrobial efficacy. In vitro findings were mostly confirmed. For drug-organism combinations which showed a predominantly time-dependent killing pattern and absence of a PAE (beta-lactams vs Gram-negative organisms), frequent drug dosing was most efficacious. In contrast, a fast, predominantly concentration-dependent bactericidal effect followed by a PAE in vitro (e.g. aminoglycosides vs Gram-negative bacteria) correlated (though inconsistently) with superiority of bolus dosing over more continuous drug administration in vivo. Thus, the ratio of peak serum concentrations to MICs of target pathogens is possibly a valid predictor of efficacy for the aminoglycosides but not so for the activity of beta-lactam antibiotics where the duration of coverage at supra-MIC levels was clearly more important than the magnitude by which initial peaks exceeded the MIC of the target organism. It is not clear to what extent the results obtained in experimental soft tissue infections may hold true in man. Thus far, only a limited number of drug-organism combinations have been studied in well defined experimental settings using mostly small, granulocytopenic animals which differ pharmacokinetically from man. In addition, results are probably affected by the density of bacteria, their growth rate and metabolic activity, but also by the extent of inflammation at the site of infection.