Clinical islet allotransplantation has been a safe, but largely unsuccessful enterprise. It has been difficult to apply techniques that might overcome the islet yield and allograft rejection problems encountered in animal experiments. Over the past decade only 4 of 74 attempts at islet transplantation have been followed by long term withdrawal of exogenous insulin therapy, and there are problems with intrepretation of the outcome in each of theses cases, as discussed in the preceding section. In the islet allograft situation, the failures may have been for technical or for immunological reasons. In the autograft situation, rejection could not occur and the failures were clearly technical. The success rate with islet autografts gives some indication as to what might be achieved with islet allotransplantation if rejection could be prevented in the latter situation. The islet autotransplant experience is not entirely predictative, however, for two reasons: 1) the uncertainty over the contribution of the pancreatic remnant to carbohydrate metabolism when less than the total pancreatectomy is done; 2) the increased difficulty with liberating islets from diseased, fibrotic pancreases. For both islet allo- and autotransplantation, the success rate will probably remain low until more effective techniques are developed for preparation of islets from adult pancreases. For the allograft situation, additional advances will be needed in immunosuppression or in techniques to alter islet graft immunogenicity in order to overcome the rejection phenomenon.