AIMS/HYPOTHESIS: It is well documented that melatonin influences insulin secretion mediated by G-protein-coupled melatonin receptor isoforms MT1 and MT2, which are present in rat and human pancreatic islets, as well as in rat insulinoma cells. Recent investigations have proven that hyperinsulinaemic Goto-Kakizaki (GK) rats, which are a rat model of type 2 diabetic rats, and humans have decreased melatonin plasma levels, whereas a streptozotocin-induced rat model of diabetes developed reduced insulin levels combined with increased melatonin levels. METHODS: Plasma levels of glucose, insulin and melatonin as well as RNA expression of pineal Aanat, Hiomt (also known as Asmt), insulin receptor, adrenoceptor β1 and the clock genes Per1 and Bmal1 (also known as Arntl) were determined in male and female LEW.1AR1-iddm rats as well as in insulin-substituted LEW.1AR1-iddm rats. RESULTS: Severe hypoinsulinaemia in diabetic LEW.1AR1-iddm rats was associated with decreased body weight and increased melatonin plasma levels combined with mainly elevated expression of Aanat, Hiomt, pineal insulin receptor and adrenoceptor β1. The changes were normalised by insulin substitution. Diurnal profiles of plasma melatonin and of antagonistic clock genes Per1 and Bmal1 were maintained in diabetic and insulin-substituted rats. CONCLUSIONS/INTERPRETATION: The assumed causal relation between elevated melatonin and reduced insulin levels in LEW.1AR1-iddm rats is supported by the observation that insulin substitution normalised these changes. Further support for this interpretation comes from the observation that in GK rats an increase of plasma insulin was combined with a decrease of plasma noradrenaline (norepinephrine), the most important activator of melatonin synthesis. These relationships between the noradrenergic and insulin pathway support the existence of melatonin-insulin antagonism.