Islets of Langerhans isolated from rat pancreas were incubated at 37 degrees C(95% O(2)/5% CO(2)) in buffered medium containing 1.0 mg/ml glucose and leucine (3)H for 1 hr (1st hr), washed, and incubated for an additional hr (2nd hr) in low glucose medium (0.5-1.0 mg/ml) containing unlabeled leucine. A portion of the islets was then extracted with acid-ethanol and the remainder were transferred to medium containing 3.0 mg/ml glucose and incubated for 2 hr (3rd and 4th hr) at 37 degrees C. The medium was exchanged at 30-min intervals and portions of the islets were extracted at the 3rd and 4th hr. The total amounts and specific activities of the proinsulin and insulin in the islet extracts and medium samples were determined after fractionation on Biogel P-30 columns in 3 M acetic acid. Maximal release of newly synthesized insulin occurred between the 3rd and 4th hr of incubation, confirming the results of Howell and Taylor (Biochem. J.102: 922. 1967). The high glucose medium increased the secretion of insulin approximately three to fourfold. The ratio of the specific activities of the insulin in the medium to that in the islets was about 1/1 during incubation in low glucose, but it increased to 2.5/1 during incubation with high glucose. The peak occurred at the 3rd hr, i.e., 1 hr after exposure to high glucose. The ratio of labeled proinsulin to insulin was slightly lower in the medium than in the islets. Addition of sufficient cycloheximide after the 1st hr to inhibit protein synthesis did not inhibit these responses. The specific activity of the proinsulin in the medium was about the same as that in the islets, and both were about 10-fold higher than the specific activity of the insulin. High glucose did not alter the proinsulin specific activity, which tended to decline throughout the period of observation. With cycloheximide present, the concentration of proinsulin in the islets steadily declined while the specific activity of proinsulin remained high, indicating that the proinsulin pool is small and is turning over rapidly. In terms both of amount and radioactivity proinsulin amounted to 6-7% on a molar basis of the insulin in both the medium and the islets. Addition of dibutyryl cyclic 3',5'-adenosine monophosphate (DBCAMP) (0.002 M) with high glucose during the postlabeling period slightly increased the rate of insulin secretion (133% of control) but did not significantly alter the other parameters. The results suggest that while newly synthesized insulin and proinsulin may be preferentially secreted to a slight degree, about 90% of the insulin released during 3 hr in response to glucose, or to glucose and DBCAMP, is derived from pre-existing granule stores. There were no indications of the existence of independent or nongranule pathways of insulin or proinsulin secretion.