In orchidectomized rats, unilateral intrapituitary implants of micropellets containing 85% testosterone propionate (TP), 85% dihydrotestosterone benzoate (DHTB), or 2½( or 5% 7α-methyl-19-nortestosterone acetate (7αMe) acted to; 1) reverse castration cell development and 2) increase intrapituitary FSH/LH ratios (either by lowering LH or by increasing pituitary FSH concentration, or both). The effects were localized: histological changes were restricted to regions contiguous to the implants, changes in LH and FSH were more pronounced ipsilaterally than contralaterally, and accessory sex organs (seminal vesicles and ventral prostate glands) remained completely or virtually unstimulated. Similarly placed micropellets containing cholesterol or progesterone had no significant effect. Subcutaneous implantation of 3, 6, or 9 androgenic pellets failed to produce intrapituitary changes, even if enough androgen was released to stimulate the sex accessories. Subcutaneous injection of larger amounts of these androgenic steroids, which strongly stimulated the sex accessories, did lower pituitary LH and increase pituitary FSH. Although some of these effects (suppression of pituitary castration cells and LH content) can be produced by estrogens, neither estrogen nor progesterone has been found to increase pituitary FSH stores. Consequently, the total pattern of intrapituitary effects of these steroids cannot be attributed to the formation of estrogenic or progestational conversion products, even though 2 of them (TP and 7αMe) can be aromatized (to form estrogens) and do exhibit antiestrogenic (progestational) activity. These findings provide the first unequivocal proof that androgens can exert feedback actions directly at the pituitary level in vivo. They therefore qualify the pituitary as a possible locus for physiological feedback of male sex hormone. (Endocrinology93: 1398, 1973)