Ten children with XO gonadal dysgenesis and ten control siblings (CS) had sequential IV tolbutamide and IM glucagon tests to ascertain serum and salivary insulin concentrations, to confirm the presence of parotid salivary insulin and to determine if these concentrations were of diagnostic value in the diagnosis of insulin deficiency. After tolbutamide, peak serum insulin concentrations were lower in the patients with Turner's syndrome (TS) than in control siblings (58 ± 10 vs 90 ± 15 μU/ml) and fractional areas under insulin curves were significantly lower in the patients with Turner's syndrome at 10 to 1 5 minutes (TS: 240 ± 16 μU-min/ml; CS: 340 ± 46 μU-min/ml, P<0.05) and at 15 to 30 minutes (TS: 562 ± 62 μU-min/ml; CS: 884 ± 128 μU-min/ml, P<0.05). After glucagon, peak serum insulin concentrations were significantly lower in Turner's syndrome than in control siblings (P<0.02, at 45 minutes) and fractional areas under insulin curves were also lower in TS than in siblings at 30 to 45 minutes (TS: 1,062 ± 185 μU-min/ml; CS: 2,180 ± 402 μU-min/ml, P<0.02). Basal salivary immunoreactive insulin (IRI) concentrations were similar in both groups: TS: 4.8 ± 2.1 μU/min; CS: 2.1 ± 0.4 μU/min. Peak salivary IRI concentrations after tolbutamide were 13.8 ± 4.7 μU/min in Turner's syndrome and 8.8 ± 1.8 μU/ml in control siblings. Peak salivary IRI values in Turner's syndrome and in control siblings after glucagon were 26.8 ± 10.1 and 13.4 ± 2.1 μU/min, respectively. While significant differences in insulin secretion in serum were detected in the two patient groups, no differences were noted between groups when salivary insulin concentrations were compared. These data confirm serum insulin deficiency in gonadal dysgenesis, the presence of immunoreactive insulin in parotid saliva, and suggest the possibility that extrapancreatic insulin synthesis could occur in the parotid gland.