The effect of donor islet mass on the metabolic normalization in streptozotocin-diabetic rats was examined by comparing normal control rats with two transplant groups receiving 800 or 2,000 islets intraportally, i.e. grafts containing respectively, 0.8 or 2 million Beta cells at 70% purity. After transplantation, basal glycaemia and daily urine volumes normalized in all recipients and remained normal in both groups over the 20-week follow-up period. After intragastric glucose challenge, the 800-islet group exceeded the normal range of glycaemia at all tested time points (0-120 min) whereas the 2,000-islet group exhibited higher glycaemia only at 15 and 30 min; no deterioration in glucose tolerance occurred during the 20-week follow-up. Both transplant groups presented lower serum fructosamine levels than diabetic control rats, normal levels being maintained only in the 2000-islet group. At post-transplantation week 21, both recipient groups exhibited the same body weight as age-matched normal control rats, but their serum triglyceride levels were significantly higher. Hepatic insulin contents were comparable to the insulin contents of the grafts at the time of implantation, representing 16 or 40% of the pancreatic insulin content in the age-matched control rats. Although islet grafts with 0.8 million Beta cells can restore basal glycaemia in adult streptozotocin-diabetic rats, implants of 2 million Beta cells are necessary to correct other parameters of glucose homeostasis. The present data indicate the need of determining the number of grafted Beta cells in studies on the metabolic outcome of islet cell transplantation. DNA content and percentage of Beta cells in the graft are proposed as useful parameters.