A stereological study of the numbers of germ cells in various stages of spermatogenesis was undertaken in testosterone-treated intact and hypophysectomized (HPX) rats. Adult Sprague-Dawley rats were given testosterone by Silastic implants, which either inhibited (3-cm length) or partially maintained (10 cm) spermatogenesis over a 13-week period. The numbers of nuclei of the various germ cell categories (spermatogonia, spermatocytes, and round spermatids) in the testes were estimated by profile counting and measurement of nuclear diameter. The numbers of elongated spermatids were determined separately in testicular homogenates. Testis weight, seminiferous tubule volume, and tubule diameter were significantly decreased in intact rats with 3- and 10-cm testosterone implants and in HPX rats, although they were partially maintained in groups with 10-cm implants compared to those in groups with 3-cm implants (P less than 0.05). The effect of 3-cm testosterone implants in the intact group was to suppress the number of spermatogonia to 57%, reduce the conversion of spermatogonia to spermatocytes to 85%, and reduce the conversion of round to elongated spermatids to 19% of the control value. This latter effect was largely overcome with 10-cm testosterone implants. In HPX rats, only 10-cm implants were effective in maintaining the conversion of round spermatids to elongated spermatids. However, testosterone alone was less effective in maintaining the conversion of spermatocytes to round spermatids, suggesting that a pituitary factor, probably FSH, was involved. It is concluded that testosterone has a major effect on the conversion of round to elongated spermatids. The conversion of spermatogonia to spermatocytes and the conversion of spermatocytes to round spermatids depend on the synergistic action of both FSH and testosterone. However, the effect of FSH is greatest on the conversion of spermatocytes to spermatids, i.e. meiosis.