In normal men of varying degrees of adiposity, alphaadrenergic blockade (one-hour intravenous infusion with phentolamine, 0.5 mg./min.) produced a 31 ± 23 per cent (mean ± S.D.) increase (n = 6, p < .02) over basal insulin levels whereas beta-adrenergic blockade (one-hour intravenous infusion with propranolol, 0.08 mg./min.) produced a 38 ± 15 per cent decrease (n = 6, p < .002) from basal insulin levels; basal glucose did not change during either type of blockade. Epinephrine (two-hour intravenous infusion, 6 /tg./min.) produced a 46 ± 8 per cent decrement in basal insulin by 7 ± 4 min. (n = 12, p < .001). This was followed by a period of recovery during which insulin levels rose to control levels by 46 ± 26 minutes and to 166 ± 36 per cent of control (p < .001) by two hours. The nadir of the decrement correlated, with a high degree of statistical significance, with basal insulin values (r = .99, p < .001), and the amount of time required during the recovery phase for insulin levels to rise to control levels correlated highly with the nadir of the decrement (r = .83, p < .001). The epinephrine-induced decrement from basal insulin did not occur in three subjects pretreated with phentolamine but did occur in four subjects pretreated with propranolol. In the four subjects pretreated with propranolol, the mean recovery time during epinephrine infusion was seventy-two minutes, compared to thirty-six minutes for the group receiving epinephrine alone. In four separate studies, propranolol was superimposed after two hours of a three-hour epinephrine infusion had elapsed, and the elevated insulin levels that had developed by two hours promptly fell below basal levels. It is concluded that (1) endogenous adrenergic activity modulates basal insulin secretion; (2) increases in circulating epinephrine prompt a rapid, marked fall in basal insulin which is alpha mediated; and (3) a slower, beta-mediated recovery phase follows during which insulin surpasses basal levels.