A large number of ascitic fluid tests, e.g., fibronectin and cholesterol, have been proposed as helpful in detecting malignancy as the cause of ascites. Unfortunately, these "humoral tests of malignancy" are nonspecific. Although the ascitic fluid concentrations of these proteins or protein-bound substances tend to be quite high in patients with peritoneal carcinomatosis and low in the setting of cirrhotic ascites, the problem is that patients with tuberculous peritonitis, cardiac ascites, pancreatitis ascites, etc. usually have values in the malignancy range, i.e., false-positive results. This can lead to an extensive search for a nonexistent tumor, with confusion and anxiety for patient and physician. The cytology is the single best test to order when peritoneal carcinomatosis is suspected; its sensitivity approaches 100%. However, peritoneal carcinomatosis is only one of several mechanisms by which tumors can cause ascites. No one test can be expected to detect tumors as the cause of these diverse mechanisms of ascites formation. The serum-ascites albumin gradient is a helpful test in classifying ascitic fluid specimens into portal-hypertension-related and non-portal-hypertension-related categories. An elevated serum alpha-fetoprotein test can be useful in raising suspicion of hepatocellular carcinoma. Careful analysis of ascitic fluid, without measurement of "humoral tests of malignancy," combined with information obtained from the history and physical examination, usually lead to an accurate diagnosis of the cause of ascites.