Pretreatment with n-butanol (10 mmol/kg i.p.) 30 minutes before alloxan (100 mg/kg) protects mice from the permanent hyperglycemic effects (measured at 72 hours) of the diabetogenic agent. This dose of n-butanol caused an elevation of serum glucose at 30 minutes, the time of alloxan administration. Since glucose administration can protect animals from alloxan, the possibility that alcohol-induced hyperglycemia protected mice from alloxan was investigated. Mannoheptulose, an antagonist of glucose action at the pancreatic beta-cell, when given 24 minutes after n-butanol and 6 minutes before alloxan, eliminated the alcohol-induced protection. Fasted mice did not exhibit n-butanol-induced hyperglycemia at 30 minutes and alloxan given at that time produced diabetes. No protection was observed in fed animals when n-butanol was given 5 minutes before alloxan. The high serum levels of butanol and normal serum glucose which were observed at 5 minutes after alcohol administration indicated that the lack of protection was not due to a lack of circulating alcohol but resulted from an absence of hyperglycemia. The results indicate that pretreatment with n-butanol protects mice from alloxan-induced diabetes by the indirect mechanism of producing hyperglycemia at the time of alloxan administration.