52 patients with serious urinary tract infections were randomised to receive either aztreonam (35) or gentamicin (17). In the aztreonam group 23 patients had unqualified cures, 6 cures with relapse, and 6 cures with reinfection; the comparable numbers in the gentamicin group were 9, 1, and 4. There were no failures with aztreonam and 3 with gentamicin. The most important determinant of outcome was the presence or absence of urological abnormalities. 11 further patients, with renal failure or gentamicin-resistant isolates, treated with aztreonam were all cured. Toxic effects were limited to symptomless liver-function-test abnormalities with aztreonam , whereas deterioration in renal function occurred in 4 gentamicin-treated subjects. Urinary colonisation with group D streptococci occurred in 14 of 46 aztreonam -treated patients (1 required treatment) compared with only 1 of 17 gentamicin-treated patients. 97% of 309 consecutive gram-negative urinary isolates tested, including 50 Pseudomonas aeruginosa, were susceptible in vitro to aztreonam and 91% to gentamicin. Aztreonam may prove an effective and safe alternative to the aminoglycosides.