Alloxan is used to induce diabetes in animals; however, the underlying mechanisms are still a matter of debate. Alloxan evoked a rapid hyperpolarization of the plasma membrane potential and suppressed electrical activity elicited by 15mM glucose, thus terminating voltage-dependent Ca²⁺ influx. Accordingly, glucose-induced oscillations in intracellular free Ca²⁺ concentration were abolished. The effect of alloxan on membrane potential could not be reversed by glucose but was reversed by tolbutamide. However, the sensitivity to tolbutamide was decreased after treatment of the cells with alloxan. These effects closely resemble those described earlier for H₂O₂. H₂O₂ and alloxan decreased the mitochondrial membrane potential, indicating a decrease in ATP production and thus interference with cell metabolism. A decrease in ATP synthesis would explain the plasma membrane hyperpolarization observed in intact islets, reflecting the activation of ATP-dependent K⁺ channels. Surprisingly, alloxan inhibited the whole-cell K⁺_ATP current measured in single cells and the single-channel K⁺_ATP current registered in excised patches. This inhibitory effect of alloxan is not mediated by changes in cell metabolism but seems to be due to direct interactions with the K⁺_ATP channels via thiol-group oxidation. We have monitored the appearance of reactive oxygen species in single cells and islets treated with alloxan and H₂O₂ for comparison. In contrast to H₂O₂, alloxan induced the appearance of measurable reactive oxygen species only in islets but not in single cells. The results show that alloxan evokes different effects in islets and single cells, giving a possible explanation for inconsistent results reported in the past. It is concluded that alloxan exerts its diabetogenic effect by the production of H₂O₂ in intact islets.