In some respects, the action of testosterone on all of its targets may be similar. Thus, the behavioral responses and that of the accessory gland diminish with increasingly prolonged testosterone deprivation. Comparison of the amount of testosterone required to reduce plasma LH in our "maintenance" experiment (Fig. 10) with the doses used by others in "restoration" experiments (41) suggests that a similar situation may obtain for the feedback mechanism.

Both the latency to loss of function after castration and that to its restoration on testosterone replacement are longer for sex behavior than for accessory gland stimulation. The feedback system has not as yet been studied in this respect, although one would certainly expect more rapid changes than are shown in See PDF for Figure the behavioral response. With regard to responsiveness to estrogen, the feedback and behavioral systems respond qualitatively as does testosterone, although estrogen is more effective on the feedback mechanism and less effective on behavior than is testosterone. A different relationship also holds for the antiandrogenic effects of cyproterone. This antiandrogen appears to interact with both the behavioral and the feedback substrates in the hypothalamus, but its action on the two receptors is different. In this case, the responses of the feedback and reproductive tract systems resemble one another and both differ from the sexual behavior mechanism.

Sexual Differentiation

No discussion of this subject would be complete without some mention of sexual differentiation, the neuroendocrine aspects of which have been widely studied in recent years. The accepted view today is that testosterone secreted by the neonatal rat testis is responsible for fixing both the acyclic pattern of male gonadotropic function and male behavioral characteristics; these effects are presumably mediated by the action of androgen on critical brain areas (47, 73, 96). It is important to recognize, however, that the pattern of female sex behavior is not completely suppressed in the male rat.

Recent studies in this laboratory have shown that estrogen, when administered in high daily doses for a week or more, is capable of eliciting patterns of lordosis in the castrate male which resemble in frequency those found in the female, albeit with much higher doses of estrogen (27). As exemplified by the data in Fig. 14, the difference between males and females in this respect is primarily one of relative sensitivity to estrogen, but the inherent capacity to respond with the feminine pattern remains present in male animals.

One qualitative distinction which we have, however, found between males and females is See PDF for Figure See PDF for Figure that males do not show the typical facilitation of lordosis when progesterone is added to estrogen. In Fig. 15, it is shown that progesterone did not change the lordosis-to-mount ratio in males chronically treated with estrogen, when the comparison was on the same day before and after treatment with progesterone. The ability to respond to progesterone is present, however, in neonatally castrated males (Fig. 16). It appears, then, that the behaviorally differentiating action of neonatal testosterone manifests itself in two ways: The lordosis response becomes relatively refractory to estrogen and the propensity to be facilitated by progesterone does not develop.

It is of interest to extend our previous comparisons of the peripheral action of testosterone to its dual central actions—on gonadotropin regulation and sex behavior—into the area of early differentiation. One obvious difference exists between these peripheral and central actions in rats. The reproductive tract See PDF for Figure and external genitalia are differentiated in the prenatal period (17), with the partial exception of the penis whose development is permanently impaired in neonatally castrated males (8). Suppression of the lordosis response is believed to be mainly a function of postnatal exposure to androgen, since estrogen-induced lordosis in the adult is suppressed after neonatal administration of testerone to females, but present in neonatally castrated males (47).

Several lines of evidence suggest, however, that the period for differentiation of sexual behavior in rats does extend partly into the prenatal period: (a) When testosterone is administered in high doses to pregnant rats, the female offspring do show some masculinization of their sex behavior (43). (b) Less testosterone is required to "remasculinize" the neonatal castrate male in terms of its future behavior than is required to "feminize" the neonatal female (72). (c) Neonatally castrated male rats show more mounting behavior than do normal females, although this could be variously interpreted (98). All these findings suggest that the differentiating effect of testosterone on mechanisms organizing sex behavior differ from their effects on organization of gonadotropic patterns in that they are not entirely limited to the postnatal period.

In one other respect, at least, these two differentiating functions of testosterone diverge, i.e., the effective circulating level of testosterone needed to produce the effects in the neonatal period are not the same. Thus, it has been shown that a lower dose of testosterone, injected in the first days of life, is required to precipitate persistent vaginal estrus and anovulation than to suppress female sex behavior patterns (72a). It would be most interesting to know whether the loci of these two differentiating actions of testosterone on the brain are different, but no information is as yet available on this question.