The effect of intrauterine transplantation (IU group) as a potential immunologically privileged site on the diabetic state of the recipient was compared with that of conventional intraperitoneal transplantation (IP group) using Fisher 344 rats. Islets were isolated from the pancreata of normal rats and transplanted into the uterus and peritoneal cavity of the isogenic rats with experimental diabetes, which were treated with estradiol benzoate and progesterone. Although all the rats in both groups became normoglycemic within 4 days after transplantation, all of those in the IU group relapsed into a diabetic state up to the 20th day after transplantation. On the other hand, 6 of 8 rats in the IP group remained normoglycemic throughout the experimental period. Weight gain and diminution of urinary glucose excretion in the IU group were significantly lower than those in the IP group (P < 0.01). The glycosylated hemoglobin level in the IU group did not differ significantly from that in the IP group, but the serum level of fructosamine in the IU group was significantly higher than that in the IP group (P < 0.01). These results indicate that the response to fluctuations of blood glucose of islets in the uterine cavity is less than that of islets in the peritoneal cavity. Histologically, islets were observed to be aggregated in the uterine cavity, however the number of cells decreased markedly with time. Although this study demonstrated that blood glucose was normalized by transplantation of islets into the uterine cavity of diabetic rats, long-term survival of the islets in this location was not obtained.