That growth hormone (GH) regulates its own secretion by negative feedback both directly and indirectly via somatomedins has been well-documented in the rat and assumed, on the basis of limited studies, to be true for man as well. Prior to proceeding with studies designed to investigate the nature of this feedback in various endocrine states, we sought first to confirm the existence of direct, short-latency GH feedback in normal individuals. Primed, continuous rate infusions of human GH in normal volunteers achieved a range of steady state GH levels. After 1 h of GH infusion, arginine HC1 (500 mg/kg) was infused over 30 min and the GH response assessed. Seven of 8 subjects achieved steady state GH levels in the 8–21 ng/ml range with an infusion rate of 0.0045 U/min following a 0.277 U bolus. After arginine, there was a significant increment of GH levels (range 11.6–69.5 ng/ml) in all 7 subjects. With a higher infusion rate of 0.009 U/min following a 0.54 U bolus, 1 of 2 subjects reached a steady state of 31.0 ng/ml and no response to arginine was demonstrable. Two subjects reached steady state levels of 45.0 and 58 ng/ml during a 0.018 U/min infusion after a 1.08 U bolus and had increments of 26.2 and 25.9 ng/ml following arginine. In 3 subjects achieving levels of 64.5, 107.0 and 132.0 ng/ml, there were increments of 55.0, 61.7 and 13.0 ng/ml during infusions of 0.036 U/min following boluses of 2.16 U. However, only in the first of these 3 was a true steady state achieved. Additionally, in these last 3 subjects, on a separate day, the primed continuous rate GH infusion was extended for 6 h prior to arginine infusion to try to stimulate the generation of somatomedin-C. These 3 subjects achieved steady state GH levels of 94.4 ±5.4, 108.5 ±8.5 and 125.9 ±5.8 ng/ml, and they had GH increments of 43.6, 21.8 and 7.1 ng/ml, respectively. Somatomedin-C levels did not change during the infusion. In all 14 GH infusions in which a steady state was reached, only 2 subjects did not demonstrate GH responses to arginine. In the 22 control arginine studies in these 11 subjects, 3 had absent GH responses to arginine. The numbers of absent responses in these control versus experimental conditions are not statistically different. We, therefore, were unable to detect significant blunting of the GH response to arginine and are unable to confirm previous reports of direct GH feedback