Pleural effusions from 100 unselected patients were studied cytogenetically by direct preparation, 24- and 48-hour cultures, and long-term cultures. The study included benign pleural effusion in 17 patients, malignant pleural effusion in 76, and an indeterminate cause in 7. Two cytogenetic criteria for detecting a malignant lesion were tested: (1) three or more metaphases that were either hyperdiploid or had markers and (2) two or more metaphases with the same markers (clone). Among the malignant pleural effusions, 51 (67%) were positive by at least one method when the first criterion was used and 34 (45%) were positive by at least one method when the second criterion was used. By the first criterion, at least one of the culture methods was positive in 13 cases (17%) in which direct preparations were negative; there was 7 such cases (9%) by the second criterion. Apparently, culture methods alone are not better than direct preparations but are useful in supplementing direct preparations. There was one false-positive long-term culture by the first criterion. Thus, the clone criterion of malignancy may be more accurate but is less sensitive. There was no evidence that therapy had any effect on the cytogenetic studies.