A close coupling in time-course and dose-dependency exists between cyclic AMP and insulin responses to glucose as tested in isolated islets of Langerhans from the rate and other species. Other secretagogues are also capable of inducing a cyclic AMP response. Under circumstances where the insulin response is reduced, secretion can, however, proceed without a measurable increase in cyclic AMP. "Classical" cyclic AMP-raising agents such as the methylxanthines are unable to induce substantial insulin release in the absence of glucose; however, short-term previous exposure to 27.7 mM glucose transforms 3-isobutyl-1-methylxanthine into a potent insulin-releasing agent. It is concluded that 1) all secretagogues share the stimulation of cyclic AMP as part of the insulinogenic signal; 2) a cyclic AMP response may be needed only for optimal secretory responsiveness of the B-cell; 3) whether a rise in cyclic AMP is solely sufficient to trigger insulin secretion depends on the extent to which a "memory" of exposure to glucose is present in the B-cell.