One of the main barriers to widespread application of islet transplantation is the limited availability of human pancreatic islets. The reduction of graft islet mass for transplantation to a recipient is one of the strategies in islet transplantation. However, transplantation of only a small number of islets may result in primary nonfunction. To optimize the sites and numbers of islets for transplantation, we analyzed these factors using pancreatic islets from Lewis or F344 rats transplanted into rats rendered diabetic by streptozotocin (50 mg/kg IV) and confirmed as such prior to transplantation (>300 mg/dL blood glucose). Approximately 500 to 1500 islets were injected via the portal vein or under the renal capsule into the diabetic F344 rats. The blood glucose level of all animals bearing 1500 syngeneic or allogeneic islets transplanted to the liver or under the kidney capsule exhibited restored normoglycemia (<200 mg/dL) at 1 day after transplantation. Graft function deteriorated after only 3 days in three animals (5.8%). The loss of graft function after 3 days occurred in 10 of 28 rats transplanted with 1000 to 1200 syngeneic islets, 4 of 19 rats transplanted with 800 to 900 syngeneic islets, and 7 of 17 rats transplanted with 500 to 600 syngeneic islets. There was no significant difference in the loss of graft function between the sites of transplantation via portal vein or under the kidney capsule. In conclusion, higher frequencies of primary nonfunction occurred with less than 1500 islets transplanted. They were independent of the sites in the rat-islet transplantation model.