A major problem that limits the effectiveness of clinical trials of islet transplantation is the inability to isolate sufficient pure viable islets from a single pancreas. The authors have evaluated this problem, using dogs rendered diabetic by total pancreatectomy. An average of 123 +/- 11 X 10(3) highly purified islets (mean graft weight 0.75 +/- 0.1 g) were implanted as splenic allografts (seven dogs), splenic autografts (six dogs) or liver autografts (six). In the autograft recipients, fasting normoglycemia was maintained during follow-up to 10 months; onset of hyperglycemia was delayed in three liver recipients and one splenic autograft recipient at 1.5, 2, 8 and 10 months respectively. The K values (decline in glucose levels in %/min) during intravenous glucose tolerance testing were more than 1.0. Six recipients of allografts and cyclosporine (CsA) were normoglycemic when CsA trough serum levels were greater than 300 micrograms/L, although the fasting plasma glucose level was higher than that in autograft recipients. These data show that purified islets isolated from a large mammal will maintain fasting normoglycemia for prolonged periods after auto- or alloimplantation.