The diagnosis and treatment of 20 hospital patients seen in the past year with proven pneumonia caused by coliforms and Pseudomonas aeruginosa are discussed. Predisposing factors and methods for improving laboratory and clinical diagnosis are analysed, the main problem being to discriminate between genuine pneumonia caused by these organisms and mere contamination of sputum samples resulting from colonization of the upper respiratory tract following broad-spectrum chemotherapy. Overall initial chemotherapy with gentamicin cured 75% (15 out of 20) of the patients in spite of unfavourable underlying pathology. Where gentamicin was given in adequate dosage, which in practice meant that dose which produced peak serum concentrations of 8 mug/ml or more, the cure rate was 91% (11 out of 12). In those patients achieving (measured) peak serum concentrations of less than 8 mug/ml the cure rate was only 33% (4 out of 12). These figures include four patients who failed to respond to doses of gentamicin producing peak concentrations of 5-0-6-0 mug/ml in each case. These patients responded promptly to higher doses (or accumulation), producing peak serum concentrations of 8 mug/ml or more and were then cured within three to five days. Toxicity from gentamicin was not observed in any patient. These results indicate that it is necessary to monitor gentamicin therapy by laboratory assay to ensure adequate dosage and that peak serum concentrations of 8 mug/ml or more are significantly correlated with successful treatment of pneumonia caused by coliforms and Ps. aeruginosa.