Since provocative tests for GH release may not reflect normal GH secretory patterns, we studied 18 children with diverse growth disorders, age 7-18 yrs, by two standard stimuli (L-Dopa, arginine,insulin-induced hypoglycemia) and by continuous blood withdrawal (collected hourly) for 12 and 24 hr integrated GH concentrations (ICGH). Tests were started at 0800 hr; 12 hr ICGH was calculated for 1800-0600 hrs. Six patients with normal stimulation tests (peak > 10 ng/ml) had 12 hr ICGH = 5.2 ± 2.6 ng/ml (mean ± S.D.; range 2.5-9.1); two boys with delayed adolescence had no hourly peak > 10 ng/ml. Five patients with borderline stimulation tests (peak GH 7-10 ng/ml) had 12 hr ICGH 4.0 ± 2.3 (range 1.5-6.7); in 3/5, the peak integrated hourly levels were low: 4.5, 4.7, 8.7. Seven patients with low stimulation tests (peak < 7 ng/ml) had 12 hr ICGH 0.75 ± 0.6 ng/ml (range 0.18-1.96); all had hourly peak ICGH < 3.5 ng/ml and 6/7 had clinical evidence of GH deficiency. Patients with delayed puberty had borderline or normal provocative GH; however, 4/5 had low 12 hr ICGH and hourly peaks. There were highly significant correlations between ICGH peak, 12 hr ICGH and peak provocative GH (p < 0.001). In all groups, 12 hr ICGH correlated with 24 hr ICGH (p < 0.001). With the exception of one child with obesity, growth velocities correlated more closely with ICGH than with provocative GH. Constant withdrawal may be helpful particularly in patients with borderline provocative tests and/or sexual immaturity.