The hypoglycemic effect of i.v. insulin injected at 4 different times of day was studied in normal volunteers. It was found that the glucose fall induced by this treatment was quantitatively different, with the maximum effect observed at 18(00) and 24(00). The circadian variation of the insulin-induced hypoglycemia depends on the time recovery from the glucose fall rather than on lower circulating glucose levels attained following hormone injection. These results underline the importance of considering circadian variations when planning either experimental designs or insulin therapy.