The diabetogenic action of the beta-cell toxin, alloxan, is transient when administered to mice at a dosage of 50 mg/kg. We examined whether increased cholinergic activity is involved in the compensatory mechanisms. Therefore, following administration of alloxan, methylatropine (32 mumol/kg) was given intraperitoneally once daily for 5 consecutive days. Methyl atropine worsened the degree of hyperglycaemia during the first week after alloxan administration. Recovery from the diabetes mellitus was observed in a substantial number of animals given alloxan without methyl atropine, whereas the risk of developing manifest diabetes was markedly enhanced by methyl atropine. At 35 days after alloxan administration, 33% of the animals, which were given alloxan alone and were diabetic after 4 days, still had diabetes. In contrast, of the animals rendered diabetic by alloxan with concomitant atropinization, 92% remained diabetic throughout the study (p = 0.0145 vs alloxan alone). Glucose-stimulated insulin secretion and pancreatic insulin content were markedly reduced in animals with diabetes while being less reduced in alloxan-injected animals without diabetes. Moreover, in situ hybridization and immunocytochemistry revealed markedly decreased levels of insulin mRNA and number of insulin cells in alloxan-treated animals. With regard to insulin secretion, pancreatic insulin content, insulin mRNA and insulin cell number, the reduction was the same irrespective of whether methyl atropine had been given. Thus, 5 days of atropinization increases the incidence of diabetes following alloxan at 50 mg/kg in mice. We suggest that cholinergic activity protects insulin cells from glucotoxicity during the first week after alloxan administration and therefore, reduces the frequency of diabetes.