Islet amyloid polypeptide (IAPP), or amylin, was originally discovered as the constituent peptide in amyloid occurring in human insulinomas and in pancreatic islets in human subjects with Type II (non-insulin-dependent) diabetes mellitus. Its normal expression in beta cells and its co-secretion with insulin in response to nutrient stimuli, suggest a metabolic function for the peptide. Specifically, IAPP has most frequently been shown to inhibit insulin secretion, implying that IAPP has a role in the regulation of islet hormone homeostasis. The physiological significance of IAPP in islets has been difficult to assess; very high IAPP concentrations are required to alter insulin secretion. Moreover, until recently, IAPP receptors have not been characterised at the molecular level, thus leaving the actual target cells for IAPP unidentified. Furthermore, in experimental diabetes in rodents, the ratio of IAPP expression to that of insulin invariably is increased. In view of the pleiotropic effects attributed to IAPP, such regulation could be both adverse and beneficial in diabetes. Metabolic characterisation of mice carrying a null mutation in the IAPP gene or which overexpress IAPP in beta cells have recently confirmed that IAPP is a physiological inhibitor of insulin secretion. Based on experiments in which IAPP-deficient mice develop a more severe form of alloxan-induced diabetes, we argue that the action of IAPP in the islets normally is beneficial for beta-cell function and survival; thus, the established up regulation of IAPP expression compared with that of insulin in experimental rodent diabetes could serve to protect islets under metabolically challenging circumstances. [Diabetologia (2000) 43: 687–695]