T o determine if both phases of glucagon secretion are excessive in diabetes, arginine was admimistered intravenously as pulses and as infusions to normal subjects, insulin-dependent diabetics, and noninsulin-requiring diabetics. The acute phase of glucagon secretion, in response to arginine pulses at four different doses (submaximal to maximal alpha-cell stimulating), was indistinguishable in terms of timing, peak levels attained, and total increments comparing controls and diabetics. During the first half of the arginine infusion (500 mg/kg over 30 min) the glucagon rise in controls and diabetics was similar (P greater than 0.1), whereas during the last half of the infusion excessive glucagon levels were seen in the diabetics. No difference in the glucagon responses to arginine administered as either a pulse or an infusion was observed between the two types of diabetics. The acute phase responses of insulin to intravenous, maximal stimulating doses of glucose (20 g) and arginine (2.5 g) were measured in five insulin-independent diabetics. Although the acute insulin response to arginine was normal, there was marked attentuation of the early beta-cell response upon stimulation by glucose. From these results we conclude that although in diabetes excessive glucagon levels are observed with chronic arginine stimulation, the acute phase of glucagon secretion in response to arginine is normal. In addition, the beta-cell in noninsulin-requiring diabetics, although acutely hyporesponsive to glucose, remains normally responsive to another stimulus, arginine.