The relative efficacy of different aminoglycosides or of different dosage schedules of the same aminoglycoside should be quantitated and related to relative toxicity. Quantitative experimental indicators of efficacy should not only include MIC and MBC, but also the postantibiotic effect in vitro and in vivo, the emergence of resistance in in-vitro and in-vivo models, and the relationship between plasma concentration profiles and efficacy. Parameters of clinical efficacy are to be related to pharmacokinetic parameters such as the ratio between the peak serum concentration and the MIC. Toxicity in clinical trials should be assessed by the most sensitive methods available. Experimental and clinical studies have shown cortical uptake to be a sensitive indicator of renal toxicity. As far as ototoxicity is concerned endolymph and perilymph pharmacokinetics are not clearly related. Clinical ototoxicity should be assessed by sensitive methods, such as high frequency tone audiometry. Finally, risk factors for nephrotoxicity and ototoxicity (e.g., duration of treatment, associated nephrotoxic drugs, dehydration) should be assessed in the evaluation of clinical trials.