Spontaneous bacterial peritonitis is an infection of the ascitic fluid of patients who, in general, have severe chronic liver disease. Several variants of this disease exist including bacterascites, culture-negative neutrocytic ascites, and secondary bacterial peritonitis. Spontaneous bacterial peritonitis is frequently manifested by signs and symptoms of peritonitis although the findings may be subtle; however, occasionally it may be completely without clinical manifestation. The clinician must have a high index of suspicion in order to make this diagnosis at a relatively earlier stage of infection. An abdominal paracentesis is required to make the diagnosis of spontaneous bacterial peritonitis. This paracentesis should be performed on all patients who are admitted to the hospital for ascites and should be repeated if there is any manifestation of bacterial infection during the hospitalization. Patients with severe intrahepatic shunting--as manifested by marked redistribution of activity from the liver to the spleen and to the bone marrow on liver-spleen scan as well as patients with an ascitic fluid total protein concentration of less than 1 g/dl--appear to be particularly susceptible to bacterial infection of their ascites. In order to optimize the yield of ascitic fluid culture, it is probably appropriate to inject blood culture bottles with ascites at the bedside immediately after the abdominal paracentesis. The mortality of spontaneous bacterial peritonitis continues to be very high. Perhaps routine admission paracentesis and prompt empiric antibiotic therapy with a third-generation cephalosporin will decrease the mortality of this infection if the Gram stain of the ascitic fluid demonstrates bacteria or the ascitic fluid neutrophil count is greater than 250 cells/cu mm. Repeating the paracentesis after 48 hours of treatment to reculture the fluid and reassess the ascitic fluid neutrophil count appears to be the best way to assess efficacy of treatment. After 48 hours of treatment the ascitic fluid neutrophil count should be less than 50% of the original value if the antimicrobial therapy is appropriate. The optimal duration of antibiotic treatment is unknown; however, until controlled trials provide data regarding duration of treatment it is appropriate to treat with parenteral antibiotics for 10 to 14 days. Research is also needed to determine if there are measures which can be taken to prevent the development of spontaneous peritonitis.