The preventive effect of ursodesoxycholic acid on pancreatic injury by alloxan (alloxan diabetes) has been reported by Watari, et al. (1976). In the following experiment, to pursue the findings further, ursodesoxycholic acid was used curatively for alloxan diabetes. A first group of animals (5 mice) were injected with alloxan (4 mg) twice at the fifth and tenth day. The second group (5 mice) was injected with ursodesoxycholic acid (0.2 mg each) for 14 days during the experiment in addition to the same alloxan dosage/frequency as the first group. A third group of animals (5 mice) served as the control. The animals were sacrificed on the 15th day and the blood sugar levels were examined, using commercial test paper. The pancreatic tissues were fixed in a mixture of 2.5% glutaraldehyde and 2% osmid acid solution, which was adjusted at pH 7.4 with a veronal acetate buffer; the osmotic pressure was also regulated by adding sucrose of 0.045 g/ml. Following dehydration using a series of alcohol concentrations, the tissues were embedded in Epon 812. Thin sections were cut with a Porter-Blum MT-2B ultramicrotome, stained with both uranyl acetate and a lead mixture, and then observed by electron microscopy. The results were as follows: The pancreatic islet cells, especially of B-cells in the first group of animals injected with alloxan only, were seriously damaged and contained myelinated mitochondria. Golgi apparatus, and an increasing number of autophagic vacuoles. Some B-cells revealed hydropic degeneration. Some B-granules changed into vacuoles after diacrine secretion. Pancreatic A-cells were increased in number and showed no cell injuries. On the other hand, the pancreatic B-cells of mice treated with both alloxan and ursodesoxycholic acid maintained almost normal fine structures. In summary, ursodesoxycholic acid has a curative effect on alloxan-induced pancreatic B-cell injury.