Urinary growth hormone (u-GH) excretion was measured in 547 healthy children and 110 adults by ELISA with a detection limit of 1.1 ng/l u-GH after prior concentration of the urine samples (20- to 30-fold). u-GH excretion values were significantly dependent on the pubertal stage (p < 0.0001) with maximum values in Tanner stage 3 for girls and 4 for boys. This corresponded to a peak in u-GH excretion between 11.5–14.5 years in girls and 12.5–16 years in boys. Additionally, u-GH excretion in adults was significantly higher than in prepubertal children (p < 0.001). The day/night ratio of u-GH excretion (pg/h) was significantly higher in females than in males (p < 0.01). In Tanner stages 1–4, u-GH excretion during the day was lower than that at night, whereas the opposite was true in late puberty and in adult women. The interindividual variation of u-GH excretion within the same Tanner stage was considerable and approximately double the intraindividual variation. The day-to-day variation could be further reduced by collection of three consecutive urine samples. The variations were larger if night samples instead of 24-hour samples were considered. The expression of u-GH excretion in nanograms per gram creatinine did not diminish the observed variation and blunted the pubertal increase in u-GH excretion. In conclusion, (1) u-GH excretion depends significantly on age, sex and pubertal maturation as does the day/night ratio of u-GH excretion. (2) The intraindividual variation in u-GH excretion is considerable. In contrast, the interindividual variation is only half as large and can be further reduced by consecutive samples. (3) It is of no advantage to relate u-GH excretion to creatinine output. Thus, we recommend that reference intervals for u-GH excretion be expressed in nano- or picograms per period. They should be based on a large number of healthy volunteers of both sexes and on either 24-hour or on several overnight urine collections.