The present survey of aminoglycoside nephro- and ototoxicity covers approximately 10,000 patients reported on in clinical trials published between 1975 and 1982. Included in the survey were clinical trials with at least 15 patients evaluable for nephro and/or ototoxicity provided relevant data were given on methodology, patient material and aminoglycoside dosage. Each publication was evaluated by both investigators and relevant data entered into a chart. One hundred and forty-four published trials were surveyed; 139, 63 and 34 for renal, cochlear and vestibular side effects, respectively. Frequencies were calculated as number of patients with side effect of total number of evaluated patients. In the average overall figures toxicity labelled by respective authors as 'definitely', 'probably' and 'possibly' related to study drug is included. When available, frequencies of toxicity 'definitely' and 'probably' related to study drug were analysed separately. The average daily dosages of gentamicin, tobramycin, netilmicin and amikacin were 3.9, 3.8, 5.2 and 15.4 mg/kg, respectively. The average frequencies of nephrotoxicity for gentamicin and tobramycin were 14.0 and 12.9%, respectively, and of netilmicin and amikacin 8.7 and 9.4%, respectively. The average frequency of cochlear toxicity was 13.9% for amikacin, 8.3 and 6.1% for gentamicin and tobramycin, respectively, and 2.4% for netilmicin. The material available for evaluation of vestibular toxicity was considerably smaller. The average frequencies for gentamicin, tobramycin and amikacin were similar (3.2 to 3.7%) while netilmicin again exhibited a somewhat lower figure (1.4%). The overall figures and the clinical ranking that they imply were basically substantiated when prospective comparative trials were analysed separately. However, some inconsistencies go unexplained: for example the frequency of gentamicin nephrotoxicity was markedly higher in trials where it was compared to tobramycin (20 trials) than when it was compared to netilmicin (16 trials).