Studies of therapy for experimental pneumonia due to Pseudomonas aeruginosa have failed to document beta-lactam-aminoglycoside synergy for most antibiotics examined, in contrast to results usually observed with pseudomonas infections at other sites. The neutropenic guinea-pig model of pseudomonas pneumonia was modified to resemble more closely therapy for clinical infections. Animals were treated 16 hr after infection with ticarcillin, azlocillin, ceftazidime, tobramycin, and netilmicin, alone and in combination. As predicted by in vitro synergy testing, in all cases combination drug therapy was more effective than the corresponding drugs given alone (P less than .05), as assessed by quantitative lung culture. Among single-drug regimens, those in which peak antibiotic levels did not exceed the minimal bactericidal concentration for the organism were significantly less effective. Resistance to aminoglycosides did not develop during therapy, and therefore, in this study does not explain the mechanism of synergy observed with beta-lactam antibiotics.