The use of an initial bolus of insulin prior to the initiation of low-dose insulin infusion therapy was evaluated in 56 episodes of diabetic ketoacidosis (DKA) in 38 children. The cases were randomly assigned to a group that received a bolus of insulin (n = 24) and to a group that did not (n = 32). After the first hour of insulin therapy, the decline in serum glucose level and the changes in serum osmolality were statistically similar in the two groups, regardless of the degree of acidosis. The time required to reach a serum glucose level of less than 250 mg/dl and the total duration of insulin infusion likewise were similar in the two groups. The use of a bolus of insulin at the onset of treatment for DKA appears unnecessary.