A 33-year-old man was admitted to our hospital in October, 1995, because of high fever for 2 months and abdominal pain and diarrhoea for 1 month. His temperature fluctuated between 37·5° and 38·5°C. He passed pasty stools with no pus or blood one to three times per day with some relief of his lower abdominal pain. Blood tests showed leucocytosis with a high ratio of eosinophils. Suspected of infection with the liver fluke Clonorchis sinensis at another hospital, he was given praziquantel tablets (which were manufactured in 1974 and were going brown) 0·5 g three times a day for 3 days without effect. At a second hospital 2 weeks previously, he had been given several antibiotics without relief. Stool examinations were negative. All the routine tests of liver and renal function and of serum electrolytes were normal. Blood and stool cultures were negative; a skin test (1/8000 lung-fluke antigen) was also negative. No microfilaria or Plasmodium vivax were found in his blood. ELISA for schistosomiasis was positive; intradermal test for schistosomiasis was weakly positive. Air-contrast barium enema showed colitis. Abdominal ultrasonography showed diffuse liver disease and computed tomography scan showed several homogeneous low-density patchy shadows in the liver and slight enlargement of the pancreas. Bone-marrow aspiration showed hyperproliferation of granulocytes with eosinophilia. On physical examination at our hospital, he was thin and pale, without rashes or jaundice. There was a lymph node about 1·5X1·0 cm in his left supraclavicular area and another, slightly smaller, in his left groin. Heart rate was 96/min with regular rhythm. Chest radiograph was normal. His liver and spleen were enlarged. Blood count showed Hb 8·2 g/dL, WBC 19X109/L, and eosinophils 0·25X109/L. Several stool examinations were negative; blood and stool cultures were negative. Bone-marrow aspiration showed eosinophilia. A biopsy specimen of the liver showed hepatocellular degeneration and necrosis with infiltration of eosinophils, but no ova or parasites. He had no allergic history and admitted to no abnormal sexual practices. He had eaten uncooked bile and blood of a viper, Agkistrodon acutus1 Huang M Qn Y Agkistrodon acutus. Beijing Science Press (Chinese), Beijing1983: 64 Google Scholar , twice in 1994 and five times in 1995. In July, 1995, he took 40 mL bile and 10–20 mL blood. He had eaten cooked frog meat in 1994, but had never eaten rats or birds.