317 patients with suspected or documented infections other than cystitis were randomly assigned to receive gentamicin or tobramycin dosed according to the Sawchuk/Zaske method or a modification of the McHenry method. 196 patients completed 6 days of therapy, had serial determinations of serum creatinine and were evaluated for nephrotoxicity. Within each dosing method group those patients receiving gentamicin and tobramycin had a similar mean age, initial serum creatinine level, total dose, duration of therapy and trough serum aminoglycoside concentration. Nephrotoxicity developed in 5 of 49 (10.2%) given gentamicin via the McHenry method compared to 9 of 49 (18.4%) given tobramycin. Similarily, 4 of 50 (8.0%) given gentamicin via the Sawchuk/Zaske method developed nephrotoxicity compared to 8 of 48 (16.7%) given tobramycin. Within the Sawchuk/Zaske and the modified McHenry dosing method groups, no significant difference was noted in the incidence of nephrotoxicity associated with gentamicin and tobramycin. 34% of patients with elevated trough serum aminoglycoside concentrations developed nephrotoxicity compared to 3.7% of those with nonelevated troughs (p less than 0.0005).