Assuming that the serum-to-saliva transfer of insulin reflects internalization and re-cycling of the hormone in the membrane-located binding sites of salivary epithelial cells and that these cells have in obesity a'marked decrease in insulin receptor content, it has been postulated that insulin resistance in infantile obesity can be detected by the changes in the salivary immunoreactive insulin during the oral glucose tolerance test (OGTT). The study included 31 obese children and adolescents of both sexes, subjected to OGTT. Samples of blood and saliva were collected at 30, 60, 120, 180 and 240 minutes for determinations of glucose and IRI. The blood glucose values were generally normal whereas IRI was excessively high. The dynamics of salivary IRI was similar (easy peak followed by slow descent) with the mean serum values but lower by about two-thirds, and the peak was 30-60 minutes delayed. The serum IRI values correlated significantly with the saliva ones at all time-intervals except for the 30-minute ones. The serum IRI values were significantly lower at the 30-minute time interval, whereas the salivary IRI were the lowest (and of borderline significance) at the 60-min. time interval. The mean glucose/kg doses given orally were not significantly different in the two groups. It was concluded that a hormonal activity detectable by IRI assay through the PEG separation method does exist, with a concomitant variation of serum-to-saliva transfer as shown by the OGTT test. It was also concluded that since the salivary values are lower, the direction of the flow is from serum to saliva and not the reverse. Finally, on the basis of our data, an "in situ" synthesis of insulin (hormonogenic exocrinism) can not be ruled out.