To evaluate gonadotropin release in polycystic ovary syndrome (PCO), one or more of the following hypothalamic-pituitary function tests were performed on 24 patients with the syndrome. These tests included (a) the pulsatile pattern and day-to-day fluctuation of gonadotropin release; (b) effects of exogenous estrogen and antiestrogen (clomiphene) administration on gonadotropin release; and (c) pituitary responsiveness to maximal (150 mug) and submaximal (10 mug) luteinizing hormone-releasing factor (LRF) injections. In 10 of the 14 patients sampled frequently (15 min) for 6 h, luteinizing hormone (LH) levels were elevated above the concentration seen in normal cycling women (except the LH surge). These high LH concentrations appeared to be maintained by and temporally related to the presence of exaggerated pulsatile LH release, either in the form of enhanced amplitude or increased frequency. In all subjects, levels of follicle-stimulating hormone (FSH) were low or low normal, and a pulsatile pattern was not discernible. In four patients, daily sampling revealed marked day-to-day fluctuation of LH but not FSH. That the elevated LH levels were not related to a defect in the negative-feedback effect of estrogen was suggested by the appropriate fall of LH in four patients given an acute intravenous infusion of 17beta-estradiol. This infusion had no effect on FSH levels. In addition, clomiphene elicited rises of both LH and FSH that were comparable to the ones observed in normal women given the same treatment. The clomiphene study also suggested that the positive-feed-back mechanism of estrogen on LH release was intact when the preovulatory rises of 17beta-estradiol induced appropriate LH surges. The elevated LH levels appeared to be related to a heightened pituitary responsiveness to the LRF. This was found in the 11 and 2 patients given maximal (150 mug) and submaximal (10 mug) doses of LRF, respectively. The augmented pituitary sensitivity for LH release correlated with the basal levels of both estrone (P less than 0.025) and 17beta-estradiol (P less than 0.02). The net increase in FSH was significantly greater (P less than 0.001) in the PCO patients than the normal women with maximal doses of LRF. With the smaller dose study none of the injections had a discernible effect on FSH concentrations in either subject. The disparity between LH and FSH secretion could be explained by the preferential inhibitory action of estrogen on FSH release, coupled with a relative insensitivity of FSH release. These data indicate that in these PCO patients the abnormalities of the hypothalamic-pituitary regulation of gonadotropin secretion was not an inherent defect but represented a functional derangement consequent to inappropriate estrogen feedback, which led to a vicious cycle of chronic anovulation and inappropriate gonadotropin secretion.