Fifty-one adult male outpatients at or below ideal body weight (IBW) with no history of weighing more than 15% over IBW in the past 5 yr and no more than 25% in the past 15 yr were randomly assigned to a traditional calorically defined exchange-type diet (EXCH) or an unmeasured diet emphasizing avoidance of refined sugar and balance of food consumption throughout the day. All but 4 patients were insulin treated. With the exception of one patient in each group, all patients were classified as having type II diabetes. Subjects were followed in a single-blind design for 4 yr in the Diabetes Outpatient Clinic every 3 mo. The average mean fasting glycemia (FBS), coefficient of FBS variation, number of reported hypoglycemic reactions, mean fasting serum triglyceride and cholesterol levels, and mean total daily insulin dosage were similar in both groups. Each patient's mean daily caloric intake did vary over time, but there was no difference in mean caloric intake between diet groups. There was a significant correlation between a patient's mean FBS rating and his serum triglyceride level: patients with lower mean FBS had a significantly lower mean triglyceride level than patients with higher FBS. Overall body weights remained stable throughout the study. However, there was a significant difference in the number of patients in the EXCH group whose actual weight was 3% or more above IBW on 50% or more of their visits. The unmeasured diet group had a significantly greater number of patients whose actual weight was at IBW during the study. There was no relationship between FBS and changes from IBW. In the EXCH group only subjects who increased their weight over initial levels had lower FBS values than patients who reduced from initial weight. It is concluded that individuals without history of obesity before the diagnosis of diabetes mellitus can continue to regulate their weight without a precise caloric prescription. All of the parameters studied suggest that a calorically defined exchange diet might be as effective as an unmeasured diet in the conventional long-term treatment of the lean adult diabetic patient.