Many advances characterize the research into the diagnosis of short stature in children. Increasing evidence shows a continuous spectrum of growth hormone (GH) output among GH-deficient patients and short normal children. Although biosynthetic human GH could theoretically offer the chance of treating most slowly growing children, it is not certain that all short normal children with a poor height velocity could benefit from therapy. Indeed, besides auxological findings, the assessment of GH secretion remains essential in selecting candidates for therapy. In this respect the evaluation of GH secretion by means of tests that can explore the physiological pathways involved in the hormone output appears important. Moreover some clinical evidence suggests that pharmacological stimuli cause the pituitary release of stored GH perhaps unavailable in physiological conditions. Among the classical physiological tests, the exercise test, the standardization of which has been debated, is commonly used in clinical practice. The sleep test, i.e. the evaluation of sleep-associated GH secretion, is the most important. It has no side effects and does not require the administration of exogenous stimuli. Several studies have demonstrated its reliability in diagnosing growth disorders in childhood, mainly if performed with EEG monitoring. Among the new physiological diagnostic approaches the most reliable test is the evaluation of 24-hour GH secretion. Knowledge of the integrated hormone concentrations appears particularly important in studying children who may have more subtle disturbances in GH secretion. These cases show normal GH response to provocative stimuli but show a reduced hormone output over 24 h. Indeed they respond well to human GH treatment.