Four hypotheses have been posited on the role of insulin in glucose-stimulated insulin secretion; available evidence has supported insulin as being 1) essential, 2) a positive modulator, 3) a negative modulator, or 4) not necessary. Because circulating insulin levels in mice, before or after intraperitoneal glucose injection, are sufficient to elicit insulin responses in insulinsensitive tissues, it is likely that -cell insulin receptors are continuously exposed to stimulating concentrations of insulin. To determine whether constitutively secreted insulin is necessary for glucose-stimulated insulin secretion, CD1 male mouse islets were incubated for 30 min at 4°C in the absence (control) or presence of anti-insulin (1 g/ml) or anti-IgG (1 g/ml). Then islets were exposed to 3, 11, or 25 mmol/l glucose or to 20 mmol/l arginine. Nontreated islets exhibited first-and second-phase glucose-stimulated insulin secretion. Control and anti-IgG-treated islets, after a 5-min lag phase, increased their insulin secretion in 25 mmol/l glucose. Anti-insulin؊treated islets secreted insulin at a basal rate in 3 or 25 mmol/l glucose buffers. Insulin secretion stimulated by 20 mmol/l arginine was the same in islets pretreated with either antibody and showed no lag phase. Taken together, these data suggest that constitutively secreted insulin is required and sufficient for -cells to maintain sensitivity to glucose.