A prospective clinical microbiological study of pleural fluid samples was conducted to investigate the etiology of pleural effusions and to evaluate two different methods for transport and culture of these samples. A total of 245 pleural fluid specimens were inoculated into a transport vial, an aerobic and an anaerobic blood culture vial, and a sterile tube. One hundred nine samples were from infectious patients and 128 from noninfectious patients. Gram stain had a sensitivity of 48% and a specificity of 100% as compared to culture. Of the total, 15.5% of the samples were positive for microorganisms, and 60% of the positive samples were nonpurulent pleural fluid. Single-organism growth was found in 23 samples (60.5%). Sixty-three microorganisms were isolated: 25 (39.7%) aerobic, 22 (35%) anaerobic, 13 (20.6%) mycobacteria, and three (4.7%) fungi. Of the 25 positive samples, excluding those samples that grew mycobacteria, nine (36%) were positive exclusively in the blood culture vials. Twelve organisms were isolated, only one of which did not grow in the anaerobic vial. Two (8%) samples were positive by conventional culture only, and 14 (56%) were positive by both methods. The microorganism isolation rate obtained with use of blood culture vials was significantly greater than that obtained with the conventional method of transport and culture. Sixty-three percent of the empyema patients had an associated underlying pathology, pneumonia being the most frequent. In conclusion, for microbiological study of pleural fluid, it seems appropriate to inoculate all samples, including nonpurulent samples, into both a sterile tube and an anaerobic blood culture vial.