Oral glucose tolerance tests (OGTTs) were conducted in 100 normal, non-obese children aged 4 to 15 years with a negative family history for diabetes. Glucose was administered in a 20% solution (1.75 g/kg body weight), and blood glucose levels were determined by the o-toluidine technique (capillary microglycemia in whole blood). The following results were obtained: fasting: 79±26 mg/100 ml ( $$\bar x$$ ±2 SD); 30 min: 121±54 mg/100 ml; 60 min: 103 ± 42 mg/100 ml; 90 min: 98±36 mg/100 ml; 120 min: 90±36 mg/100 ml. Individually, 70 children showed peak blood glucose at 30 min (sub-group A), 19 at 60 min (sub-group B); 5 at 90 min (sub-group C), and 6 showed flat curves (sub-group D). The $$\bar x$$ +2 SD values of some of these sub-groups slightly exceeded that of the whole 100-children group. This suggests that normal curves pertaining to these sub-groups might be considered pathologic if referred to the statistical values of the whole groups. Out of the 100 children, 4 showed one value above $$\bar x$$ +2 SD, and 3 showed two values above that limit. At 120 min, however, none of the individual or group values were above $$\bar x$$ ±2 SD (126 mg/100 ml), which supports other authors’ criteria regarding the importance of the 120-min value in this test. On the basis of the present findings, normal OGTT’s were defined as those in which all values except for one, which must not be the 120-min value, fall within the $$\bar x$$ ±2 SD range calculated for the whole group. Normal children may present two values, except for the 120-min value, slightly above $$\bar x$$ +2 SD, although we consider retesting advisable in such cases.