The reliability of provocative stimuli of GH secretion in the diagnosis of GH deficiency is still controversial. Until now, normative values of GH response to various stimuli have not been established properly. In 472 children and adolescents with normal stature (n = 295, height SDS range -1.5 to 1.2) or normal short stature (n = 177, height SDS range -3.7 to -1.8), we studied the GH response to physical exercise, insulin-induced hypoglycemia, arginine (ARG), clonidine, levodopa, glucagon, pyridostigmine (PD), GHRH, PD + GHRH, and ARG + GHRH. The peak GH responses (range) to various stimuli were: 1) physical exercise: 3.0-28.3 micrograms/L; 2) insulin-induced hypoglycemia: 2.7-46.4 micrograms/L; 3) ARG: 0.5-48.4 micrograms/L; 4) clonidine: 3.8-86.0 micrograms/L; 5) levodopa: 1.9-40.0 micrograms/L; 6) glucagon: 1.9-49.5 micrograms/L; 7) PD: 2.5-35.0 micrograms/L; 8) GHRH: 2.7-102.7 micrograms/L; 9)PD + GHRH: 19.6-106.0 micrograms/L; and 10) ARG + GHRH: 19.4-120.0 micrograms/L. Our results show that all conventional stimuli of GH secretion frequently failed to increase GH levels, showing values lower than that arbitrarily assumed, so far, as minimum normal GH peak, i.e. 7 or 10 micrograms/L. When combined with PD or ARG (substances inhibiting hypothalamic somatostatin release), GHRH becomes the most powerful test to explore the secretory capacity of somatotrope cells (the GH response being always higher than 19 micrograms/L). Therefore, only GHRH combined with PD or ARG may be able to clearly differentiate normal children from patients with GH deficiency, though a normal GH response to these tests cannot rule out the existence of GH hyposecretory state because of hypothalamic dysfunction.