While the paramount role of estrogens, and sometimes also progesterone, in the control of female sexual behaviors is clearly established in most mammalian species, the situation is less clear in primates. While non-primate female mammals usually display sexual behavior only during a limited period of the estrus cycle, generally the period during which estrogen concentrations are high, female monkeys copulate at any time during the cycle (Zuckerman, 1930). The first study that established this fact was published only a few years after the first descriptions of the dramatic cyclic variations in female rat sexual behaviors. This temporal coincidence probably contributed to the now firmly established notion that the hormone dependency of female rodent sexual behavior is far superior to that found in primates. There is, undoubtedly, a substantial amount of support for this notion, but female primate sexual behavior is, as we will see, by no means independent of gonadal hormones. Nevertheless, the fact that female monkeys copulate throughout the menstrual cycle was soon confirmed in a carefully designed study (Ball and Hartman, 1935). Many later studies have described copulatory behavior throughout the primate menstrual cycle (e.g. Michael and Zumpe, 1970; Slob et al., 1978; Dixson, 2001). Some researchers report increases in some aspects of female sexual behaviors in the periovulatory period (e.g. Czaja and Bielert, 1975), while others fail to find any systematic variation during the menstrual cycle (e.g. Nikitopoulos et al., 2005).
A particularly interesting example of a primate female with almost constant sexual behaviors throughout the menstrual cycle is the capuchin monkey. At difference to the rhesus and many other primate species in which the females show evident exterior signs of the periovulatory phase, like substantial swelling of the sexual skin, the capuchin monkey has concealed ovulation. With that is understood that there is no morphological or known behavioral changes associated with ovulation. Nevertheless, the male partner displays enhanced affiliative and copu-latory behaviors during the periovulatory phase (Carnegie et al., 2005). Since neither the female's behavior nor her appearance is different at that period, the males must use some unknown sign or signs for discovering that ovulation is around. If females in a species with concealed ovulation are more attractive to males around ovulation, it is no wonder that females in species displaying exterior morphological and/or behavioral signs during that period also are more attractive than at other periods of the menstrual cycle. The often described enhanced periovulatory sexual activity in female primates with visible signs of imminent ovulation can, therefore, be entirely explained as a consequence of more intense behavior on the part of the male during this period, as suggested many years ago (Michael and Bonsall, 1977). It would appear, then, that primate female sexual behavior is little influenced by the cyclic variations in blood concentrations of ovarian hormones.
The absence of a clear-cut relationship between phase of the menstrual cycle and the female monkey's ability to display sexual behaviors raises immediately doubts as to the importance of ovarian hormones in the control of that behavior. In fact, sexual behavior persists after ovariectomy in several primate species, including the rhesus, stumptail macaque, talapoin and marmoset (Slob et al., 1978; Dixson, 2001). Some behavioral changes are frequently observed, but most of these changes may be attributed to modifications of the male's responses to ovariectomized females (Herbert, 1970). Such females are considerably less attractive to males than intact females are and this leads to reduced intensity of sexual activity. At difference to the slight effects of ovariectomy, suppression of adrenal androgen secretion with dexamethasone (Everitt and Herbert, 1971) or adrenalectomy (Everitt et al., 1972) causes a strong reduction of female sexual behavior, at least in the rhesus monkey. Treatment with estradiol benzoate fails to restore sexual behaviors in these animals, while androgens like testosterone or androstenedione readily do so. These observations suggest that androgens rather than estrogens are important for female primate sexual behavior, at least in the rhesus monkey.
In other primate species, like the stumptail macaque, it appears that female sexual behavior is entirely independent of hormones. I already mentioned that ovariec-tomy fails to affect sexual behaviors in females of this species and there are data showing that these behaviors persist undiminished after treatment of ovariec-tomized females with dexamethasone or after combined ovari- and adrenalectomy (Baum et al., 1978). Treatment with either estradiol or testosterone had little effect, which is not particularly surprising since the absence of these hormones did not reduce sexual behaviors in the first place. This carefully performed study could suggest that there are important species differences among primates with regard to the importance of androgens. While they seem to be crucial in rhesus females, they are not necessary for stumptail macaques.
Rather than using species differences as an explanation for conflicting observations, it can also be proposed that experimental results are unreliable, varying from one experiment to another. This latter idea is somewhat strengthened by the results of an annoying study in the rhesus monkey. Instead of reducing androgen availability by suppression of adrenal hormone production through treatment with dexam-ethasone or adrenalectomy as was done in the Everitt and Herbert (1971) and Everitt et al. (1972) studies, this study employed an androgen receptor antagonist. It could reasonably be predicted that the functional consequences of blocking receptors should be quite similar to those of reducing androgen availability. Unfortunately, this did not turn out to be the case. The androgen antagonist flutamide failed to affect sexual behaviors when given to intact, cycling rhesus females (Johnson and Phoenix, 1978). I called this an annoying study not because it was badly performed but because it is difficult to understand why the results are so different from those obtained in earlier studies. Despite a substantial effort, it has been impossible to find a convincing explanation. The Johnson and Phoenix (1978) experiment did confirm one thing, though. In agreement with many other studies, it was found that stage of the menstrual cycle was unrelated to the intensity of sexual behaviors.
I will not try to review all the studies that have tried to determine the importance of hormones for female, non-human primate sexual behaviors. The examples given in the preceding paragraphs should be enough to illustrate the state of confusion in which that field is submerged. To strengthen that impression it might be appropriate to mention that one prominent scientist still insists upon the existence of systematic variations in female primate sexual behavior according to phase of the estrus cycle with a maximum around ovulation. Associated with this comes also an insistence on the role of estrogens and the lack of a role for androgens. The arguments supporting his points of view have been summarized in several reviews (e.g. Wallen, 1990; Wallen, 2004). I consider myself unable to dismiss all these arguments and I feel for the moment compelled to conclude that the endocrine control of non-human primate female sexual behaviors remains mysterious.
It is impossible to leave this issue without trying to offer some kind of explanation for the current sad state of affairs. There are, in fact, several possible reasons for the confusion. The most important, I believe, is the use of small numbers of animals. Many of the primate studies mentioned here were performed on very few subjects. A number of subjects as low as four pairs is not unusual and few studies employed more than 10 pairs. The low number of subjects may have many practical justifications, the main one probably being the prohibitive costs of maintaining large primate colonies. In addition, primate work is time- and personnel-consuming. None of these arguments have any scientific validity, though. I cannot see much reason to believe that interindividual variability between primates should be any smaller than that found in rats or mice. On the contrary, I would suspect that variability among monkeys is superior to that found between rodents. In addition to a badly known genic variability, it is probably much easier to assure equality of experiences among individuals in a standard rodent colony than it is in a standard monkey colony. The shorter life span of a rodent makes it simply impossible to acquire the same amount of experience as a monkey. It may consequently be suspected that pre-experimental experiences play a larger role in primate than in rodent studies. In view of these trivial considerations and many other unmen-tioned, I doubt that primate studies a priori require fewer subjects than rodent studies. Furthermore, I can see no reason why data obtained from a primate should be more valid than data obtained from a rodent and there is no reason to assume that statistical procedures are more trustworthy when they are employed to analyze primate data than they are when used for rodent data.
To make the arguments exposed in the preceding paragraph clearer, I will give an example from my own rat data. Some time ago we were interested in determining whether a new compound could activate male rat copulatory behavior. During these studies we employed dihydrotestosterone as a kind of control. We ran several experiments, some with rats with very little sexual experience, some with rats with substantial sexual experience. For purposes of illustration, I pick four rats treated with dihydrotestosterone from a group of 12 animals with little experience and another four from a group of 12 animals with more sexual experience. I also make a random sample of four controls (untreated, castrated males) of the 12 used in each experiment. As can be seen in Figure 5.1a, it turns out that the experienced animals are far superior to the controls while the inexperienced animals are not. The conclusion would be that the effects of dihydrotestosterone on male rat sexual behavior depend on experience. If I instead include all 12 animals from each group with their respective control groups of 12 animals, there is no effect at all of dihydrotestosterone and, consequently, none of experience (Figure 5.1b). This example is certainly not perfect, but it should illustrate how random variations between individuals can affect experimental results when the number of subjects is low. As soon as we approach a more reasonable group size, the influence of exceptional subjects becomes much reduced.
Random interindividual variations in response to hormones may be the cause of much of the conflicting primate data. A wise solution to this apparent problem is simply to follow the majority. A majority of studies have reported that female
Experience No experience ;;;; DHT □ Cholesterol
Experience No experience HI DHT □ Cholesterol
Experience No experience üü DHTD Cholesterol
Experience No experience DHT □ Cholesterol
FIGURE 5.1 (a) Mean (± SE) number of mounts and intromissions in groups of four male rats each with extensive sexual experience (at least 15 tests terminated at the end of the first post-ejaculatory interval before the animals were castrated) and subcutaneously implanted with two 25 mm long Silastic® capsules containing dihydrotestosterone or cholesterol 35 days after castration (Experience) and from two other groups of four male rats each that had been exposed to sexually receptive females only twice before castration (No experience). The latter groups were also implanted with subcutaneous Silastic® capsules exactly as the experienced animals were. All four groups of four rats each were selected from treatment groups that contained 12 animals each. (b) Data from all 12 animals included in each group are illustrated in this figure. Note that the significant difference between the groups given dihydrotestosterone and cholesterol obtained with the small group sizes (illustrated in (a)) disappeared when data from more animals were added (as in (b)). Likewise, the significant difference between experienced and inexperienced rats treated with dihydrotestosterone obtained with the small group sizes (a) is no longer found when data from more animals are used (b). Data shown here are from a test performed 21 days after implantation of the Silastic® capsules. DHT, dihydrotestosterone. *, different from cholesterol according to the Mann-Whitney U-test, P < 0.05. +, different from animals with no experience (actually animals with little experience), P < 0.05. The experiment from which these data stem was performed by Xi Chu.
primate sexual behavior does not vary substantially during the menstrual cycle. We conclude, then, that it is independent of ovarian hormones. A majority of studies have found that androgens rather than estrogens are important. We conclude, then, that female primate sexual behaviors are androgen-dependent. Some may consider that this pragmatic way of reaching a conclusion is superficial. Perhaps it is, but it is at least based on some kind of rational reasoning. Furthermore, when we turn from non-human primates to women, we will find abundant support for both conclusions.
Was this article helpful?