As we noted in our introductory chapter, over four decades ago researchers offered a definitive conclusion about women’s sexuality: Over the course of the last several million years, women had lost estrous sexuality. Modern women possess no distinct fertile sexuality within their cycles. The theoretical perspective we presented in chapter 8 casts serious doubt on this conclusion. Nearly all vertebrates, we suggest, possess estrous sexuality. Many researchers, however, have misunderstood estrus. Estrous sexuality does not function to obtain sperm or ensure fertilization; these effects are by-products of estrus’s function. Female reproduction is only very rarely limited by the availability of male sperm; typically, the abundant male willingness and ability to deliver sperm that sexual selection yields ensures that females have no trouble obtaining sperm. Rather, estrus functions to lead females to target as sires males who offer genetic benefits for offspring, relative to other males. Females in estrus should not be indiscriminate in their mate choices. Indeed, they may be even choosier during estrus than outside of it. In species with extended sexuality, what should distinguish estrous sexuality from extended sexuality are mate preferences: the features that females find most sexually attractive.
Nearly a decade ago, we discussed what has become known as the ovulatory shift hypothesis (Gangestad & Thornhill, 1998; Thornhill & Gangestad, 1999b). (We should note, however, that this hypothesis was anticipated several years earlier by Karl Grammer’s [1993] pioneering work on women’s scent preferences. Hence, he should be credited for formulating this hypothesis.) The hypothesis predicts that women prefer male features that would have been indicators of male ability to deliver genetic benefits to offspring, particularly when they are fertile in their cycles. That is, when fertile in their cycles, women should find purported indicators of good genes more attractive than they do outside of the fertile phase. Penton-Voak et al. (1999) offered an important addendum to this hypothesis: Preference shifts should be particularly strong when women evaluate men’s “sexiness” or attractiveness as sex partners, not their attractiveness as long-term, pair-bonded mates. The traits women prefer in long-term partners should be relatively constant across their ovarian cycles. Hence, when fertile in their cycles, women should find purported indicators of good genes more sexually attractive than they do outside of the fertile phase.
In the current chapter, we review the considerable body of evidence that speaks to the ovulatory shift hypothesis, most of it accumulated in the past decade. In chapter 10, we address the implications of women’s estrus for an understanding of human mating patterns considered more broadly.
Women’s estrus is comparable in length to that of many other vertebrates. The “fertile window” is an approximately 6-day period of time ending on the day of ovulation. Probability of conception associated with insemination rises gradually for a couple of days, then more steeply, peaking 1 to 2 days prior to ovulation, if not the day of ovulation itself (e.g., Wilcox et al., 1995; Dunson, Baird, Wilcox, & Weinberg, 1999). The day before the window and the day after ovulation are associated with very low probabilities of conception. The timing of the window in the menstrual cycle is somewhat unpredictable, even in women with regular cycles (e.g., Dunson, Weinberg, Baird, Kesner, & Wilcox, 2001; Jöchle, 1973; Wilcox, Dunson, & Baird, 2000; Wilcox, Dunson, Weinberg, Trussell, & Baird, 2001). A common pattern, however, is that shortly after menses, which typically lasts 5 days, probability of conception rises and peaks on about day 12, then sharply declines at ovulation, typically on day 14 in a 28-day cycle (Wilcox et al., 2001; see figure 9.1). In a sample of 221 women without fertility problems and attempting to conceive, Wilcox et al. (2001) found that 84% had regular cycles. Peak fertility in an individual woman’s cycle, however, is brief, typically 2–3 days preceding ovulation. Presumably, this phase has the greatest homology with the estrous phase of most other vertebrates.
Bullivant et al. (2004) found that women initiated sex most frequently during the fertile window (though see our extended discussion of the long-debated issue of whether women’s “sexual desire” is maximal during the fertile window in chapter 10). Based on this finding, they suggested that women’s fertile window be termed “the sexual phase.” As they appropriately noted, some terminology conventionally used by human reproductive biologists is problematic. For instance, the “follicular phase” spans days that differ dramatically in terms of typical types and quantities of hormones released and associated conception risk. Nonetheless, Bullivant et al.’s term for the fertile window, the sexual phase, is perhaps even more misleading. It implies that women are not sexual outside of this phase (i.e., do not possess extended sexuality), which is plain wrong. The term estrus captures the comparative homology, as well as important components of evolved estrous function (see chapter 8 in conjunction with evidence presented in this chapter and the next) and hence, we argue, rightly applies to women’s fertile window.
Figure 9.1 The probability of pregnancy/conception risk with one act of intercourse in relation to day of the menstrual cycle for women with regular and irregular cycles. N = 221, 16% irregular cycles. Based on figure 2 of Wilcox et al. (2001); reprinted with permission of Elsevier.
Grammer (1993) assessed the ovulatory shift in women’s scent preference for a particular chemical (an androgen; see section on the scent of androgens). By contrast, we first tested the hypothesis by examining whether fertile women are particularly attracted to the scents of particular men, ones who arguably possess an (ancestral) indicator of genetic quality. Specifically, we predicted that women particularly prefer the scent of symmetric men when fertile in their cycles. As we discussed in chapter 7, low fluctuating asymmetry reflects a relatively high level of developmental stability, which, in turn, may reflect low mutation load, ability to resist pathogens, and/or ability to resist the untoward effects of ontogenetic stresses in people.
In our initial study (Gangestad & Thornhill, 1998), we recruited 42 male participants. We measured each man’s symmetry on 10 traits (see chapter 7). We then gave each man a clean T-shirt to wear for two consecutive nights during sleep. Men were instructed to wash their sheets in unscented soap (which we provided), not to sleep with another person, not to eat certain strong-smelling foods (e.g., spicy foods), not to drink alcohol or smoke, and to refrain from using scented deodorants or cologne. Men wore their shirts the same two nights and, on the morning of the third day, each man returned his shirt to us in a clear plastic bag identified by an arbitrary code number. We then had 28 normally ovulating women (each placed in a small room by herself) smell each shirt and rate its scent on dimensions of “pleasantness” and “sexiness.” These two ratings were highly correlated, and hence we added a woman’s ratings of a man’s shirt to form an overall measure of how attractive she found the scent of that shirt. For each woman, we regressed her attractiveness ratings of the shirts on men’s symmetry and took the slope (changes in ratings as a function of changes in symmetry) as a measure of a woman’s preference for the scent of symmetry.
In this study, we estimated each woman’s fertility risk based on her day of the cycle and actuarial probability of conception given by a study by Jöchle (1973). In general, fertility risk is near zero the first several days of the cycle and after day 17. It is highest in the window of days 9–13. We found a strong, statistically robust correlation between women’s probability of conception and preference for the scent of symmetry (see Gangestad & Thornhill, 1998, for details).
Subsequently, three other studies using similar double-blind methodologies have replicated this finding: Thornhill and Gangestad (1999b), Rikowski and Grammer (1999), and Thornhill et al. (2003). These studies have extended findings, as well. For instance, Thornhill and Gangestad (1999b) controlled for the number of showers that men took. (Men could bathe with unscented soap as often as they wished, and, in that study, number of showers covaried positively with the attractiveness of men’s scent.) Partialing out this factor actually increased the association between women’s fertility risk and their preference for the scent of men’s symmetry. (This procedure had this effect because the factor controlled for—the number of showers men took—was unrelated to men’s symmetry.) In addition, the effect remained when men who had been ill during the days the shirts were worn were removed from the analysis. Furthermore, women’s preferences for the scent of symmetry do not appear to be mediated by women’s greater sensitivity to men’s scent; women’s ratings of intensity of the scents as a function of symmetry did not change across the cycle. Nor did women’s ratings of intensity of the scents change across the cycle. Figure 9.2 shows how women’s preference for the scent of symmetry changes across the cycle in the three studies we conducted.
Figure 9.2 The relationship between normally ovulating women’s preferences for the body scent of symmetrical men and the women’s risk of conception across the menstrual cycle. Each point is the regression coefficient of a woman’s scent attractiveness ratings of T-shirts worn by men regressed on men’s body asymmetry. Conception risk is based on actual data from a large sample of women (Jöchle, 1973; Baker & Bellis, 1995). The line is the least squares regression line. Figured is a summary of data from three separate studies (Gangestad & Thornhill, 1998; Thornhill & Gangestad, 1999b; Thornhill et al., 2003); N = 141 women. The relationship was statistically significant in each of the three studies.
Two studies included reproductive-age women who were using hormone-based contraceptives (a contraceptive pill or Depo-Provera). These women showed no preference for the scent of men’s symmetry when mid-cycle (e.g., Gangestad & Thornhill, 1998; Thornhill & Gangestad, 1999b).
Havliček, Roberts, & Flegr (2005) examined whether young women not using hormonal contraception would prefer the scent of young men high on the trait of social dominance, as assessed by a standard self-report questionnaire. Forty-eight men wore cotton pads in their armpits for 24 hours. Freshly collected pads were presented to 65 women, who rated their sexiness. Women varied in cycle phase when they made their ratings. Women in the fertile phase of their cycles, but not women in infertile phases, rated the high-dominance men as sexier smelling than the low-dominance men. Preference for the scent of dominant men was furthermore particularly strong for fertile-phase women in a pair-bond relationship. Fertile-phase women who were single did not reveal a robust preference for the scent of dominant men (figure 9.3). Havliček et al. (2005) also had the women rate the pads for intensity of the scents. Fertile and infertile women rated the scents of nondominant men as more intense. Fertile-phase women’s attraction to the scent of male dominance does not, then, appear to be mediated by greater sensitivity to men’s odors.
Figure 9.3 Mean (± SE) correlation coefficient between the psychological dominance score of men (N = 35) and the odor attractiveness rated by single (open bars) and romantically partnered women (shaded bars) in the fertile (late follicular) and infertile phases of their cycle. The difference between the two categories of single women is not statistically significant, whereas that between the two groups of partnered women is (p < 0.001). There were 30 female scent raters in the fertile phase and 35 female raters in other cycle phases. The raters were not using hormonal contraception. Based on figure 1 of Havliček et al. (2005). Reprinted with permission of The Royal Society.
No one has established, at this time, what chemicals in men’s scents are responsible for women’s preferences for the scent of symmetric men or their preferences for dominant men (or, indeed, whether these chemicals are the same for both symmetry and dominance). Some research hints at possibilities. The androgen androstenol, a chemical precursor of androstenone, importantly contributes to body odor. It gives scent a musky odor. Its production is furthermore sexually dimorphic, with men producing much more than women (see review by Gower & Ruperelia, 1993; see also Pause, Sojka, Krauel, Fehmwolfsdorf, & Ferstl, 1996). Women appear to be more sensitive to the scent of real or synthetic musk when fertile in their cycles (see review by Grammer, 1993; also Savic, Berglund, & Lindström, 2005).
Grammer (1993) was the first researcher to explicitly test the hypothesis that women particularly prefer a putative male marker of genetic quality when fertile. He conjectured that androstenol and androstenone may reflect testosterone levels in men, which may in turn reflect genetic quality. He found that young women not using hormonal contraception find the scent of androstenone (placed on a pad in a laboratory setting) least unpleasant during days of high fertility in the menstrual cycle (days 9–12). By contrast, women using hormone-based contraception did not show this response during the same cycle days. As Grammer (1993) recognized, these findings raised doubts about the essentially universal conclusion in the literature of human reproductive biology that estrus had been lost during the evolution of woman. Hummel, Gollisch, Wildt, and Kobal (1991) had actually found the same effect of fertility on women’s preference for the scent of androstenone 2 years earlier, but they did not discuss the findings in light of function for seeking a sire of high genetic quality. We note that these researchers found that women rate the scent of androstenone “less unpleasant” when fertile. Androstenone does not generally yield a pleasing scent. By contrast, androstenol, from which it is derived, does. (Androstenol has a smell similar to that of sandalwood.) Future research may examine whether women particularly prefer the scent of androstenol when fertile.
At this time, it is not known whether symmetric or dominant men emit high levels of androstenol, such that women’s preference for this androgen in fact mediates fertile women’s preference for the scent of symmetric and dominant men. As we discussed in chapter 7, markers of masculinity and testosteronization may well be indicators of male robustness and genetic quality (or, more accurately, could have been such indicators ancestrally). And, as we discuss in this chapter, fertile women are particularly attracted to a variety of masculine traits in men. For these reasons, we surely would not be surprised if female preferences for the scent of androgens when fertile do indeed tell the story behind women’s preferences for the scent of symmetric and dominant men. Future research should address this possibility.
At the same time, researchers have not found a robust effect of cycle phase on women’s preferences for the scent of men with high concurrent testosterone levels. Based on a fairly small sample of 19 Finnish men and a sample of 36 non-pill-using women, Rantala, Eriksson, Vainikka, and Kortet (2006) found no association between scent ratings and men’s T levels during any cycle phase. (They did find evidence that women prefer the scent of men with high cortisol levels, but these preferences did not vary across the cycle.) In one of our T-shirt studies mentioned earlier, we measured T levels of men. We found a trend in the predicted direction: Women tended to prefer the scent of men with high T levels more strongly when fertile than when infertile (unpublished data). (We found no evidence of a preference for the scent of men with high cortisol levels.)
Following our first study on women’s preferences for the scent of symmetry, other researchers began examining other preferences. In particular, Penton-Voak and his colleagues explored how women’s preferences for masculine facial features change across the cycle. Through computerized technology, multiple faces can be digitized and combined to create average faces. Perrett et al. (1998) used this methodology to produce an image corresponding to an average male face and an image corresponding to an average female face. Each average face reflects average sex-typical features. Though many facial features are quite similar, on average, across men and women, some facial features differ across the sexes. These latter features are influenced during development by reproductive hormones—in particular, testosterone and estrogen (see chapter 7). By exaggerating the differences between the average man’s and woman’s faces or averaging the sex-typical features in a variety of proportions, Perrett et al. created an array of male faces that varied from relatively androgynous to hypermasculine. When asked which face they found most attractive, normally ovulating Scottish and Japanese women, on average, actually tended to choose a male face somewhat more feminine than the average male face. In other samples and in other populations, women, on average, prefer a slightly masculine face (e.g., Johnston, Hagel, Franklin, Fink, & Grammer, 2001; Penton-Voak et al., 2004; Penton-Voak & Perrett, 2000; see also chapter 7, this volume).
Penton-Voak et al. (1999) examined whether women’s preferences shift across the cycle. Specifically, they hypothesized that, no matter what face women preferred on average, when fertile normally ovulating women prefer a face more masculine than the face they most prefer when infertile. This prediction was based on the hypothesis that a masculine (e.g., more testosteronized) male face is a marker of greater genetic quality (or would have been ancestrally; see chapter 7, this volume). In addition, and as noted previously, Penton-Voak et al. (1999) predicted that this shift would be pronounced when women evaluate men’s attractiveness as short-term mates (e.g., sex partners) but not as long-term, stable mates.
In two different studies—one in the United Kingdom and one in Japan—Penton-Voak et al. (1999) found the predicted shift toward greater preference for facial masculinity during women’s fertile phase. And this result appears to be robust. Additional studies revealed precisely the same shift: one using U.K. samples (Penton Voak & Perrett, 2000) and one using a mixed U.S./Austrian sample (Johnston et al., 2001). One additional study found mixed support for the hypothesis (Scarbrough & Johnston, 2005). Relatedly, Roney and Simmons (in press) found that estrous women, compared with women in other phases, particularly find men who possess high concurrent levels of testosterone facially attractive.
Some of these studies specifically examined changes in women’s evaluation of men in short-term mating and long-term mating contexts separately. All yielded the effect only when women evaluated, in essence, men’s “sexiness” and not when women evaluated men as stable, long-term partners (Penton-Voak et al., 1999; Johnston et al., 2001).
The preference shift is apparently not a by-product of a mid-cycle shift in women’s generalized differences in face perception. Johnston et al. (2001) found no fertile-cycle-phase shifts in the other facial ratings they studied (such as women’s judgments of female attractiveness, ratings of men’s dominance based on facial features, and a variety of other ratings).
One neuropsychological study found additional evidence that women’s responses to men’s faces change across the cycle (Oliver-Rodriguez, Guan, & Johnston, 1999). The P300 is a brain response appearing approximately 300 milliseconds after the presentation of a stimulus. It generally covaries with the emotional salience of the stimulus. Oliver-Rodriguez et al. (1999) found that the magnitude of the P300 response of the evoked potential of normally ovulating women in estrus (late follicular phase) correlated with their rating of male facial attractiveness but not with their ratings of female facial beauty. During the infertile phase, women’s responses covaried with both male and female attractiveness judgments. The greater emotional salience of attractive men’s faces to women during the fertile phase of their cycles may well be related, at least in part, to enhanced response to men’s facial masculinity. Additional neuropsychological studies are needed to pinpoint the nature of these effects.
The shift in women’s preferences for masculine features at estrus may yield by-products in face perception. Macrae, Alnwick, Milne, and Schloerscheidt (2002) presented a series of male and female faces to normally ovulating women and asked them to categorize, as quickly as possible, each face as either male or female. They found that, when fertile, women could more quickly categorize male faces. Their ability to categorize female faces did not change across the cycle. One interpretation of this finding is that women are more attuned to masculine features when fertile. Macrae et al. (2002) hint that the shift reflects women’s greater need to categorize male faces quickly when fertile. As we see little obvious advantage resulting from ability to categorize masculinity just milliseconds faster when fertile, we pose the alternative interpretation that this effect on speed of categorization is in fact not functional in and of itself but rather is a by-product of women’s attunement to (and preference for) male facial masculinity.
Frost (1994) found that normally ovulating women are more attracted to the faces of men with darker skin pigmentation when they are fertile than when they are infertile. In one sense, this effect is yet another instance in which fertile women particularly prefer masculine male faces. Within all races, skin tone is sexually dimorphic: Men have browner skin tone (reflecting greater concentrations of melanin in the skin) and ruddier skin tone (reflecting greater concentrations of hemoglobin in the skin; Frost, 1994; Jones, 1996). In turn, these sex differences appear to implicate sex hormones. In particular, testosterone tends to enhance melanin production, whereas estrogen suppresses its production in a variety of species, including humans (see review in Manning, Bundred, & Henzi, 2003). Just as fertile women particularly prefer structural facial features that are masculine, then, they also appear to prefer color and/or texture of faces that are masculine.
One might wonder why testosterone promotes production of melanin in the skin, whereas estrogen suppresses it. Melanin appears to be an antimicrobial. It defends against infection by viruses, bacteria, fungi, and possibly malaria (Mackintosh, 2001; Manning, Bundred, & Henzi, 2003). It may also possess antioxidant properties (McGraw, 2005), though under some conditions it appears to contribute to oxidative stress (Hegedus, 2000). Trade-offs between costs and benefits of internal and external deposition of melanin may be key to understanding sex differences and the cue or signal value of melanin.
As we have seen, women are particularly attracted to masculine faces when fertile. Moreover, male facial masculinity may be associated with male symmetry (Gangestad & Thornhill, 2003a; but see Koehler et al., 2004). Based on these associations, one might expect that estrous women will be particularly attracted to the faces of men who possess more symmetric bodies. And, indeed, a study supports precisely this prediction: Thornhill and Gangestad (2003b) found that women’s conception risk predicts positively their preference for the faces of men whose bodies were measured to be relatively symmetric (figure 9.4).
Naturally, this result must be mediated by factors other than male bodily symmetry, as women never viewed men’s bodies. That is, men’s symmetry must have been associated, in this sample, with facial features women find particularly attractive when fertile. Though we did not measure specific facial features in this sample of men to explore which ones were responsible for the preferences, we have reason to suspect a role for male masculine facial features (see also Scheib et al., 1999).
Figure 9.4 The relationship between women’s facial attractiveness preference for men with symmetric bodies and the women’s probability of conception across the menstrual cycle (r = 0.38, p < 0.005). Each point is the regression coefficient of a woman’s physical attractiveness ratings of the men’s facial pictures regressed on the men’s body symmetry. Conception risk is based on actual data (Jöchle, 1973; Baker & Bellis, 1995). There were 65 men who were rated and 45 women (nonusers of hormone-based contraception) raters. The line is the least squares regression line. Data from Thornhill & Gangestad (2003b).
Another possibility, however, is that men with symmetric bodies also have more symmetric faces. And, perhaps, fertile women are particularly attracted to male faces that are symmetric (just as fertile women are particularly attracted to masculine faces). Koehler and colleagues (Koehler, Rhodes, & Simmons, 2002); Koehler, Rhodes, Simmons, & Zebrowitz, 2006) evaluated this hypothesis. They found that, in general, women prefer symmetric male faces over asymmetric male faces (as also found by other researchers; see Rhodes, 2006, for a review). Their studies, however, revealed no change across the cycle in women’s preferences for symmetric faces; infertile women preferred symmetric faces just as much as fertile women preferred these faces. Unfortunately, Koehler and colleagues used menstruating/early follicular phase women as the infertile group in both studies. By contrast, most studies include women in the luteal phase. As high levels of progesterone—which is elevated only during the luteal phase—may suppress fertile phase preferences and hence importantly account for variation in preferences (e.g., Jones, Little, et al., 2005; Puts, 2006b; Garver-Apgar, Gangestad, & Thornhill, in press), studies that exclude women in the luteal phase may drastically underestimate differences across phases. Studies that have included women in the luteal phase have yielded mixed results: In two studies, Little, Jones, Burt, and Perrett (2007) did find that women particularly prefer more symmetric male faces when in the fertile phase of their cycles, especially when they are pair bonded and are evaluating men as short-term sex partners, whereas Cárdenas and Harris (2007) found no effect. (We note that some studies have found no evidence of an association between facial and body symmetry; e.g., Gangestad & Thornhill, 2003a. Partly on that basis, in fact, we have expressed concern that facial symmetry may not tap organism-wide developmental instability particularly well and, hence, may not function as a cue of developmental instability; see Gangestad & Thornhill, 2003a. More work on these issues is needed.)
Women prefer more masculine faces and, it appears, more masculine scents when fertile. Perhaps not surprisingly, then, they also prefer more masculine male voices. Voice pitch is sexually dimorphic. The fundamental frequency (pitch) of men’s voices is lower than that of women’s voices. And this difference holds up even when differences in body size are statistically controlled. Of course, the pitch of boys’ voices lowers during puberty, and voice pitch becomes increasingly sexually dimorphic during adolescence. Testosterone and, perhaps, other androgens play roles in these changes. In general, male voices that women most prefer are of lower pitch (i.e., are more masculine) than average male voices (see review by Puts, 2005).
Puts (2005) experimentally manipulated men’s voices to be of greater or lesser pitch. He then had 142 normally ovulating women rate the sexual attractiveness of men for long-term and short-term relationships based on voice alone. (Through the experimental manipulations, content and phrasing of individual men’s readings could be controlled.) Women fertile in their cycles preferred lower-pitched male voices more strongly than did women infertile in their cycles. This pattern held in analyses of natural variation in men’s voices, as well as analyses of experimental variations in men’s voice pitch. Furthermore, this effect was particularly strong when women rated men’s sexiness (attractiveness in short-term relationships) rather than their attractiveness as stable mates. Put otherwise, fertile women rated low-pitched male voices more attractive for short-term sexual relationships than for long-term or committed relationships. But the same could not be said of infertile women. (See also Puts, 2006b, who found that hormonally contracepting women do not have the same preference mid-cycle.)
Feinberg et al. (2006) further replicated and importantly extended these findings. They experimentally manipulated two sexually dimorphic vocal traits: vocal pitch and variation associated with length of the vocal tract. Women in the fertile phase of their cycles preferred voices more masculine along both dimensions (i.e., deeper voices and voices emanating from longer vocal tracts) than did women in infertile phases of their cycles. Feinberg et al. (2006) furthermore demonstrated that women’s preference shifts are specific to male vocal traits. They found no evidence that women’s preferences for the two vocal traits in women’s voices change across the cycle. (Feinberg and colleagues, 2006, also found that women’s preferences for male vocal masculinity are moderated by female estrogen levels; we discuss this finding further later in the chapter.)
As we discussed in chapter 7, Simpson et al. (1999) studied men’s direct intrasexually competitive behavioral displays in a situation in which men competed for a potential lunch date with an attractive woman (e.g., men’s explicit put-downs of another man and claims that they were “better” than their competitors). Men who exhibited more of these displays, as predicted, tended to be symmetric compared with their counterparts who refrained from making direct comparisons. In a subsequent study, Gangestad and colleagues (2004) were interested in assessing normally ovulating women’s reactions to these and related displays and how their attraction to them changed across the cycle.
In Simpson et al.’s (1999) study, all 76 men had been videotaped. Gangestad et al. (2004) had these videotapes coded for a variety of specific behaviors (e.g., amount of time spent smiling, amount of time looking downward), impressions (e.g., how confident men appeared, how nervous men were), and verbal content (e.g., whether men said they were superior to their competitors, whether they emphasized that they were nice guys). Factor analysis of these ratings revealed two largely independent dimensions. First, men varied in the extent to which they displayed social presence (e.g., appeared composed vs. looked downward). Second, men varied in the extent to which they displayed intrasexual competitiveness (e.g., said they were superior to their competitors vs. reflecting the personalities of nice guys). Gangestad and colleagues (2004) then had 237 normally ovulating women view a portion of the interview (about 1 minute per male participant) of approximately half the men (with each woman rating one of two different subsets of the men) and rate each man on dimensions of attractiveness as a stable mate (long-term mate attractiveness) and attractiveness as a sex partner or affair partner (short-term mate attractiveness).
Results of the study are shown in figure 9.5. As predicted, women were particularly attracted to men who displayed both social presence and intrasexual competitiveness when they were fertile—but only when they evaluated men as short-term, and not long-term, mates. Preferences for the two traits were analyzed separately, and this pattern was found for both. Gangestad et al.’s (2004) finding that women’s preferences vary as a function of fertility only when they rate men’s sexiness and not their long-term attractiveness is one typically found in studies that separately evaluate both aspects of attractiveness. As already noted, short-term preferences, but not long-term preferences, for male masculine facial features, male voices, and male facial symmetry (in one study) have been reported. As we discuss later, women’s short-term preferences, but not long-term preferences, for men’s creative production and tallness have also been demonstrated.
Figure 9.5 Normally ovulating women’s preference for men’s social presence (SP) and direct intrasexual competitiveness (DIC) as a function of the day of the cycle and long-term versus short-term mating context; points along lines are 3-day moving averages. Preference is the average regression slope of individual women’s ratings regressed on men’s SP and DIC, with men’s physical attractiveness ratings statistically controlled. High-fertility days span from day 6 to day 14, with peak fertility at day 12. Based on figure 1 of Gangestad et al. (2004). Reprinted with permission of Blackwell.
Just as male faces and voice can reveal masculinization, so too can male bodies. Testosterone promotes muscle growth, often in the service of intrasexual competition and mating effort (see chapter 7). Women prefer male bodies that are muscular, particularly through the chest, arms, and back, without being clumsily overbuilt (e.g., Frederick & Haselton, 2007). If male muscularity, similar to other masculine features, is a marker of male condition and, ancestrally, may have been an indicator of good genes, might muscularity also be a trait that women find particularly attractive when fertile?
In their study of men videotaped for a potential lunch date, Gangestad, Garver-Apgar, Simpson, and Cousins (2007) had women rate all men for their degree of muscularity. Fertile women, compared with their infertile counterparts, did indeed find the men who were rated as relatively muscular particularly attractive as short-term, relative to longterm, mates. This effect held up when men’s behavioral displays (social presence and intrasexual competitiveness, each found to be particularly sexy to fertile women) were statistically controlled.
In a related experimental study, Little, Jones, and Burriss (2007) found that women particularly prefer men’s bodies that are masculine more when fertile than when infertile in the cycle, and this effect was stronger when women evaluated men as short-term (sex) partners than long-term relationship partners.
These results may explain findings from laboratory studies examining women’s responses to male nudity. In seminal work, Slob, Bax, Hop, Rowland, and ten Bosch (1996) found that women are physiologically more aroused by and responsive to sexually explicit visual stimuli mid-cycle. Krug, Plihal, Fehm, and Born (2000; see also Krug, Pietrowsky, Fehm, & Born, 1994) found that, when they are fertile rather than infertile in their cycles, normally ovulating women exhibit a higher event-related brain potential (a late positive component occurring about 500–700 milliseconds poststimulus) after viewing nude men than after viewing various nonsexual stimuli. This brain response purportedly reflects added processing of the stimulus and hence interest. And, indeed, fertile women rated the nudes more positively than women at menses or in the luteal phase. As the nude males shown to women likely exhibited attractive musculature (as they were intended to be sexually provocative stimuli), we suspect that these results derive from a shift toward greater interest in and sexual attraction to attractive, muscular, male bodily features when women are fertile.
Indeed, in research we recently conducted, we found convergent evidence that women are more sexually attracted to and stimulated by attractive visually presented male features when fertile than when infertile in their cycles (Gangestad, Thornhill, & Garver-Apgar, 2007a). We asked a sample of approximately 50 women (all of whom were involved in committed romantic relationships) to fill out two self-report measures twice, once when fertile (as verified by a luteinizing hormone surge, which occurs about 24–48 hours prior to ovulation) and once when infertile (typically in the midluteal phase). First, we asked about their mate preferences for physical attractiveness (with 10 items, including “I place a very high importance on a potential mate’s physical attractiveness”; “Unattractive facial features are a real turn-off to me, even if the person has other positive attributes”; and “It’s hard for me to understand why some people place such high importance on a person’s physical attractiveness” [reverse-keyed]). Second, we asked about women’s interest in attractive bodily features (with 10 items, including “I find the thought of a very attractive body of the opposite sex very exciting”; “If I met someone I found very attractive right now, I would fantasize about what they would look like without clothes on”; and “Seeing the arm or leg muscles of an attractive opposite-sex person subtly flex would be a real turn-on right now”). Women were asked to think about how they felt at that moment, not how they felt in general.
As predicted, we found ovulatory shifts in both preferences. Women reported greater preference for physical attractiveness and greater interest in attractive bodily features when fertile than during the luteal phase (see also Gangestad, Garver-Apgar, Simpson, & Cousins, 2007, on fertile-phase preferences for physical attractiveness). The fertility shift in women’s interest in attractive bodily features was especially strong in this study. Future research should specify more particularly the nature and significance of male bodily features and their displays that are particularly attractive to fertile women.
Perhaps relatedly, Pawlowski and Jasienska (2005) found that women more often preferred a greater degree of sexual dimorphism in stature (i.e., taller men) when they were in the fertile phase of their menstrual cycles and when the partners were chosen for short-term relationships.
In light of Miller’s (2000) hypothesis that creativity is among an array of “mental fitness indicators” in humans, Haselton and Miller (2006) asked whether normally ovulating women’s preference for men’s talent and creativity over their preference for men’s material success (resource holdings) is enhanced when women are fertile in their cycles, and especially when women evaluate men as short-term mates. Women read a pair of vignettes. One depicted a man who displayed artistic or entrepenurial talent but had relatively little financial holdings to show for it. Another depicted a man who had financial holdings but little artistic or entreprenurial talent. When in the fertile phase of their cycles, women favored the talented man over the wealthy man when evaluating men as short-term partners. Their preferences for one sort of man over another as long-term, stable mates did not change across the cycle.
One limitation of this finding is that women may have read into the vignettes a variety of characteristics that differentiate men. These inferred features, rather than talent versus wealth per se, may have driven the effects. Thus, for instance, women may have inferred that talented but financially less endowed men were less constrained, open, and broadly competent and confident. One or another of these traits may have caused the observed effects.
Miller (2003) has suggested that fertile women may find humorous men particularly sexy. He found that fertile women rate humorous men (as described in vignettes) more attractive than they do when infertile. The critical mating context (short term vs. long term) by humor (highly humorous vs. not humorous) interaction, however, was not significantly robust. These results must be interpreted cautiously.
As we discussed in chapter 7, men’s symmetry covaries with typical functional and anatomical cerebral lateralization. Furthermore, as mentioned, a number of studies demonstrate a positive association between men’s IQ and their developmental stability. As we previously concluded, though it is plausible that intelligence and associated talents (e.g., creativity, possibly ability to produce humorous content) are fitness indicators, they do not possess the hallmarks of other fitness indicators (e.g., being particularly preferred in short-term mates). Though Haselton and Miller’s (2006) findings suggest that these traits may be particularly preferred in short-term mates by women at mid-cycle, other evidence is inconsistent or equivocal in this regard. In their videotape-rating study, Gangestad, Garver-Apgar, Simpson, & Cousins, (2007) found no evidence that men who appeared “intelligent” were particularly preferred as short-term or long-term partners when women are fertile (for further details, see the next section). More work on shifts in women’s preferences for male mental fitness indicators is required before any firm conclusions can be drawn.
Women are sexually attracted to masculine features and a variety of other indicators of developmental stability when fertile. But is it simply the case that all of their mate preferences strengthen when fertile? The view that we have put forward—that women should be particularly attracted to ancestral indicators of men’s genetic quality during estrus—predicts no. Women have preferences for traits that are not and were not ancestrally indicators of good genes. In particular, if men engaged in parental care and men and women cooperatively raised offspring, women should have evolved preferences for male traits indicative of willingness and ability to provide care (and this is the case; see chapters 4 and 7, this volume). Indeed, as we previously noted, Buss’s (1989b) cross-cultural survey of mate preferences revealed that the characteristic most valued in a mate by both men and women across cultures was “kindness and understanding.” We see no reason to believe that men who are kind and understanding to their mates have higher genetic fitness. Indeed, as we also have noted, one study of college students involved in romantic relationships indicated that symmetric men invest less in their relationships (Gangestad & Thornhill, 1997a) and are less faithful to partners (Gangestad & Thornhill, 1997b) than asymmetric men. Compared with less masculine men, men with masculine faces are viewed as less trustworthy, faithful, and investing in their relationships (e.g., Penton-Voak & Perrett, 2001) and poorer long-term mate choices (e.g., Kruger, 2006; see also Scheib, 2001).
We have discussed women’ preferences for male displays of social presence and intrasexual competitiveness. In additional analyses of the same videotapes of men, Gangestad, Garver-Apgar, et al. (2007) had separate samples of women rate how men came off in terms of 10 different traits women might find attractive in a mate. They then examined which traits women found particularly attractive in a short-term mate when fertile (that is, the three-way interactions between men’s trait level, women’s estimated conception risk, and relationship context—short term vs. long term). When fertile, women were particularly sexually attracted to men perceived as confrontative with other men, arrogant, muscular, physically attractive, and socially respected. In contrast, no shifts in women’s preference for men perceived as intelligent, kind, likely to be financially successful, or possessing qualities of a good father were detected. And men who were perceived to be faithful were actually less sexually attractive to fertile women than infertile women. (Put otherwise, fertile women were particularly sexually attracted to men who appeared to be unfaithful types.) The traits perceived to be particularly sexy to women during their fertile phase are traits found to be more valued in short-term partners than in long-term partners. The traits that fertile women did not find particularly sexy are ones more valued in long-term mates than short-term mates—the qualities of good, stable mates. (As we previously discussed, Haselton & Miller, 2006, found that infertile-phase women prefer uncreative men with wealth over creative but poorer men. As creativity and wealth were confounded in this study, cycle effects on preferences for each cannot be independently estimated.)
Some traits valued in long-term stable partners may be particularly attractive to women with high progesterone levels, characteristic of the infertile luteal phase and pregnancy. Jones, Little, et al. (2005) examined normally ovulating women’s preferences for femininity in men’s and women’s computerized faces in relation to their progesterone and estrogen levels across the menstrual cycle, as estimated from published norms for cycle days. Hormone-level estimation came from published data based on the cycle day. When women’s progesterone levels were high, they found femininity in both men’s and women’s faces more attractive than when their progesterone levels were low. Estimated estrogen levels did not predict either preference.
Both men and women with feminine faces are perceived to be more cooperative and helpful (e.g., Perrett et al., 1998). Women who may be pregnant (during the luteal phase) or are pregnant, then, may prefer investing social partners. Relatedly, DeBruine, Jones, and Perrett (2005) found that estimated progesterone levels positively predicted women’s preferences for self-similar male and female faces. This effect may reflect a preference to ally with kin in preparation for or during pregnancy. Perhaps relatedly, Wedekind et al. (1995) found that women using hormonal contraception (which raises progesterone, as does pregnancy) prefer scents of MHC-similar scent donors over scents of MHC-dissimilar donors. Similarity of MHC alleles may have functioned, ancestrally, to mark kin relatedness.
During infertile phases of the cycle, women have other preferences for male traits that may connote potential for receipt of male-provided services and benefits. Estimated progesterone level in the cycle of normally ovulating women positively predicts their preference for men with less masculine voices (Puts, 2006b). Similarly, Frost (1994) reported evidence that women in the luteal phase (associated with high levels of progesterone) prefer less masculine facial skin tone. As we discussed in chapter 7, men’s testosteronization appears to negatively covary with their romantic and paternal investment.
Women who are using hormone-based contraception do not appear to show shifts in preferences for indicators of genetic quality, as we have noted. Hormone-based contraception typically raises progesterone levels (Gilbert, 2000) and softens the mid-cycle peak in estrogen and hence creates a hormonal milieu more typical of a woman in the luteal phase. That women using hormone-based contraception do not show cycle effects suggests that hormones do play a role in these effects. As Puts (2006b) notes, different fertile-phase preferences may be affected by different hormones or combinations of hormones (see also Garver-Apgar, Gangestad, & Thornhill, in press; Roney & Simmons, 2008; Welling et al., 2008).
Perhaps also related to women’s special concerns when pregnant are findings that progesterone levels are associated with increased attraction to men who appear healthy (and less attraction to men who appear unhealthy). If a woman contracts an infection when pregnant, the developing fetus may be harmed. Immune responses are modulated by progesterone in ways that facilitate successful implantation of the zygote. When progesterone is elevated, women tend to avoid eating foods that could contain harmful bacteria, especially meat (Flaxman & Sherman, 2000; Fessler, 2001, 2002). Jones and colleagues (Jones, Little, et al., 2005; Jones, Perrett, et al., 2005) manipulated the perceived current health in computerized composite faces (while controlling for face shape and hence facial masculinity and symmetry). To do so, they had people rate perceived health and then created separate composites of faces, one of healthy-looking faces and another of unhealthy-looking ones. Healthy-looking faces had a darker color, whereas unhealthy ones exhibited pallor. Women in the luteal phase of their cycles, pregnant women, and women using oral contraceptives show greater aversion to unhealthy faces (Jones, Perrett, et al., 2005). All of these conditions are associated with high progesterone, which may thus be responsible for the effects. Indeed, this research found that estimated progesterone levels predict women’s aversion to unhealthy faces. This shift may reflect an adaptation similar to avoidance of foods that may carry pathogens: to promote healthy fetal development by avoiding contagion.
These findings do not necessarily contradict Frost’s (1994) finding that fertile-phase women prefer masculine, darker male faces, despite the fact that healthy faces preferred by luteal-phase women appeared to have darker skin tone. Healthy and masculine skin tone may vary in a largely independent fashion, as two different dimensions. More work may be needed to identify more specifically the nature of these variations.
As we discussed in chapter 7, traits that signal superior genes do not necessarily signal current health (e.g., Getty, 2002; Kokko et al., 2002; though they may well typically covary with immunocompetence [as opposed to parasite load]; Møller, Christe, & Lux, 1999). Women’s preferences for good-genes indicators and avoidance of mates with current infections may hence be partially independent, as Jones et al. discuss. Preferences for men who are disease-free may have been selected more strongly by nongenetic material benefits that result from affiliating with others who are healthy currently (and hence capable of delivering material benefits now) and will not pass on contagious infections.
We have argued that, when fertile, women particularly prefer in mates a subset of valued traits: indicators of male genetic quality. When outside the fertile phase of their cycles, they maintain strong preferences for traits that are particularly valuable in a long-term mate. As we delve into more deeply in chapter 10, preferences for good-genes indicators when women are mid-cycle may lead some women to seek genetic benefits from men other than primary partners (i.e., extra-pair partners). Specifically, not all women can possibly have as primary partners men who possess indicators of good genes. Those women who do not may be those most likely to be attracted to men other than primary partners when fertile (see chapter 10).
Some women, however, may be able to obtain as primary partners men who possess indicators of good genes. In particular, women who have high value on the mating market may be able to attract these men as primary partners. These women, then, may benefit from pursuit of (and hence preference for) men who possess these features throughout the cycle. And, indeed, as we mentioned in chapter 7, women who are physically attractive have particularly strong preferences for male facial masculinity and facial symmetry (Little et al., 2001). Attractive women show stronger preferences than unattractive women for male facial attractiveness for long-term mateships, whereas unattractive women show stronger preferences for male attractiveness than unattractive women for short-term mateships (Little et al., 2001; Penton-Voak et al., 2003; see also Clark, 2004, and Rhodes et al., 2005). Moreover, the preferences for male facial and vocal masculinity of women who have relatively high estrogen across the cycle (and who may, then, have higher reproductive capacity than other women; see chapter 6) shift less (and hence persist at higher levels throughout the cycle) than do the preferences of women with relatively low estrogen (Feinberg et al., 2006). (For preferences for vocal masculinity, see figure 9.6.) These findings purportedly reveal strategic differences in women’s preferences, dependent on their own attractiveness to men. Specifically, they support the hypothesis that attractive women can extract more materials and services from attractive men and hence reveal more consistent sexual preference for male masculine traits and symmetry across the cycle.
These findings illustrate a more general trend for biologists to increasingly recognize that female mate selection is often condition-dependent (e.g., Lynch et al., 2005). An early example is work on alternative female choice tactics in a species of scorpionfly (Thornhill, 1984b). Females of this species appear to trade off good-genes choice for material benefits delivered by males (nuptial gifts). Both female large body size (and hence high energetic cost to maintenance) and hunger independently and positively affect the willingness of females to copulate with males with small food gifts. Such males are inferior hunters and competitors with other males and thus, apparently, are of relatively low genetic quality. Thornhill (1984b) reviewed the literature of reports of other species that imply adaptive, conditional shifts in female choice. For evidence on a variety of nonhuman species supporting the hypothesis that female choice is adaptively condition-dependent, see Lynch et al. (2005).
Figure 9.6 Normally ovulating women’s average estrogen (mean of low and high fertility phases of the menstrual cycle) negatively predicts (p = 0.02) the magnitude of the cycle shift in preference for male vocal masculinity. The masculinity preference is a woman’s preference at estrus minus her preference outside estrus (early follicular and luteal). Women with relatively low estrogen show the greatest cyclic shifts in the preference. The line is the least squares regression line. Based on figure 2 of Feinberg et al. (2006). Reprinted with permission of Elsevier.
As we discussed in chapter 7, good genes take different forms. The kind of good genes preferences discussed thus far—preferences for masculine traits and indicators of developmental stability—are purportedly preferences for intrinsic good genes. These are genes good for the offspring of any woman. Compatible or complementary good genes are genes good in a mate of an individual woman because they work well with her particular genes. MHC genes dissimilar from the woman’s own genes may be instances of compatible genes.
One question that arises is whether women should, from our theoretical perspective, be expected to prefer compatible genes, as well as intrinsic good genes, when fertile. It is not clear that they should. On the one hand, compatible genes are, just like intrinsic good genes, good genes. And if estrus promotes mating with males who possess good genes, women should prefer males who provide compatible good genes at estrus. On the other hand, compatible genes may also be sought after in long-term, stable mates. Men with intrinsic good genes should be highly sought after by many women for those genes. They hence make less stable mates. Indeed, it is simply not mathematically possible for all women to end up pair bonded to men who possess high levels of intrinsic good genes. A man who, for a given woman, possesses compatible genes, by contrast, should not be highly sought after by all other women; he will not possess genes compatible with many other women. It is quite possible, in fact, for all women to be pair bonded with men who possess compatible genes. For this reason, it may make sense for women to prefer indicators of compatible genes in the infertile phases of their cycles, as well as the fertile phase.
In one of our three T-shirt studies of preferences for scent, we collected blood and typed each individual’s MHC alleles at three loci, A, B, and DRβ. As discussed in chapter 7, Claus Wedekind and colleagues (Wedekind et al., 1995; Wedekind & Füri, 1997) found that individuals of both sexes have a sexual preference for the scent of others who possess dissimilar MHC alleles at these loci (see also Santos et al., 2005). Studies prior to ours did not examine variation in preferences across the cycle. (In fact, Wedekind selected women who were in the fertile phase only.) In our study with colleagues, we did so (Thornhill et al., 2003). We found no evidence for it; women’s fertility risk was almost totally unrelated to their preference for the scent of MHC dissimilarity (r = 0.03). (In fact, as discussed in chapter 7, we found no overall preference for MHC dissimilarity in women, though we did find one in men). The lack of cycle-related variation, again, is not inconsistent with the view that estrous sexuality includes adaptation that functions to lead women to seek men with good genes. (Indeed, as we discuss in the next chapter, MHC dissimilarity to a mate does have implications consistent with our view of estrus.)
One female preference that we did detect was a preference for the scent of men who are heterozygous at MHC loci. As we described in chapter 7, this preference was very robust. As we noted there, a main benefit of having a heterozygotic mate is that he could produce a family that is more diverse at MHC alleles than could a homozygotic mate. Ancestrally, heterozygotes may have had longer survival and greater health and hence have been better investors in offspring (Mitton, 2000). Either benefit is most valuable in a long-term partner with whom a female will have multiple offspring. The former benefit, in particular, may matter little if a female has just one offspring with the chosen male (see chapter 7).
Interestingly, then, we found little variation across the cycle in women’s preference for heterozygotic males (indeed, we found a nonsignificant trend for women to particularly prefer the scent of heterozygotic males outside of the fertile period; see figure 9.7). Moreover, the pattern of preference for heterozygosity across the cycle significantly differed from the pattern of preference for the scent of men’s symmetry in the same study (which peaks, of course, at mid-cycle). Possibly, during the fertile phase of the menstrual cycle, preferences for indicators of good genes in sex partners increase, with corresponding dampening of the strength of preferences for traits particularly valuable in longterm partners.
Figure 9.7 Women’s preferences for the scents associated with men’s MHC heterozygosity across the menstrual cycle. Points along line represent 5-day moving averages. Total number of women raters was 65. Based on data from Thornhill et al. (2003).
In addition to preferring indicators of male genetic quality when fertile in their cycles, such women should also be expected to possess adaptations that lead them to avoid poor mating options. One set of poor mating options are those that would produce maladapted offspring. Consistent with theory, then, during the fertile phase of the cycles, women exhibit greater disgust toward maladaptive sex (e.g., incest and bestiality) than they do during infertile phases. As should also be expected, they do not exhibit greater disgust toward nonsexual objects (e.g., excretory products, rotting foods; Fessler & Navarrete, 2003).
Mid-cycle, women may also act to avoid sexual coercion even more than they do when infertile (e.g., Bröder & Hohmann, 2003; Chavanne & Gallup, 1998; Petralia & Gallup, 2002). Chavanne and Gallup’s antirape theory states that: (1) ancestrally, men sometimes circumvented female choice and forced copulation with females; (2) selection led to female antirape adaptations (e.g., wariness of going alone to places where women risk rape); (3) although rape was always costly to female victims (e.g., owing to physical harm and psychological trauma), women paid higher costs of rape during the fertile phase, for at that time they paid additional costs of possibly having a child sired by a father they did not choose and who may not invest in the child; and (4) because efforts to avoid rape entail costs (e.g., by limiting movement), the effort that women put into antirape tactics has been adaptively shaped to be most intense when the benefits of risk reduction were the greatest—that is, when the costs of rape were most devastating. Women do claim that they would avoid situations perceived to be those associated with higher risk of sexual coercion when fertile more so than when infertile in their cycles (Bröder & Hohmann, 2003; Chavanne & Gallup, 1998; see also chapter 12, this volume.)
A considerable body of research reveals that the value women attribute to putative markers of male genetic quality increases during their fertile phase of the menstrual cycle. Fertile women particularly prefer the scents associated with male symmetry, social dominance, and particular androgens; male facial masculinity; male vocal masculinity; male behavioral social presence and intrasexual competitiveness; male bodily muscularity and sexual attractiveness; and, possibly, facial symmetry (though findings are mixed). Fertile women experience greater disgust in response to maladaptive sex, such as incest and bestiality, and tend to avoid situations perceived to be associated with risk of sexual coercion.
The theory that urged researchers to explore changes in women’s preferences across the cycle, leading to the discovery of these effects, states that women have been designed by past selection to manifest preferences for sires with superior genetic quality for offspring. These preferences should be maximal during the periovulatory phase because that phase coincides with maximum risk of conception and, hence, is the time at which female choice has the greatest effect on offspring genetic quality. That pursuit of high genetic quality is the fundamental design of these fertile-phase preference shifts is consistent with the robust finding that the effects are specific to women’s attraction to men’s “sexiness” and do not apply to women’s preferences for men as long-term, stable partners.
We propose that these effects reflect adaptations, ones functional in the ancestral past. They are adaptations shaped and maintained by genetic benefits ancestral females obtained for their offspring by mating with men with superior genetic quality. Some effects demonstrated in the literature are mere by-products of these adaptations (e.g., the effect on women’s ability to discriminate male faces quickly, the effect on women’s interest in nude men). The shifts in preferences for specific male features at estrus in particular reflect adaptation.
Might women’s preference shifts at estrus have been shaped by benefits other than genetic benefits for offspring? Two possibilities are that women prefer masculine men mid-cycle to obtain benefits of physical protection and a fertile ejaculate.
It is surely plausible that masculine men could better protect women from unwanted advances by other men. Though we found that symmetric men invest less into their relationships overall, they perceive themselves and are perceived by their romantic partners as better able to physically protect those partners (Gangestad & Thornhill, 1997a). Protection is a form of male investment in a relationship that perhaps minimally interferes with a man’s efforts to attract partners other than his pair-bond mate. Male protective ability may be attractive to women in general and, compared with investment in the form of time, honesty and sexual exclusivity, men’s investment in protection may compete with their pursuit of additional partners to a lesser degree. Quite possibly, female preference for male protection occurs in addition to choice for superior genes. At the same time, it is not obvious that, if females choose men for physical protection alone, they should do so specifically when fertile.
Women’s preferences for symmetric and masculine men conceivably reflect adaptation for obtaining a fertile ejaculate. Manning and colleagues (Manning, Scutt, & Lewis-Jones, 1998; Manning, Scutt, Wilson, & Lewis-Jones 1998; see also Baker, 1997; Firman, Simmons, Cummins, & Matson, 2003) found that men’s body symmetry positively predicts ejaculate size and sperm quality. Soler et al. (2003) found a positive association between men’s facial attractiveness and semen quality. The benefits of obtaining good sperm need not be alternatives to obtaining superior genes. Men’s ejaculate quality may be importantly affected by individual variation in parasite resistance and immune system quality. Skau and Folstad (2003) argue that the ability to bear the cost of the immunosuppression necessary for producing high-quality ejaculates is possessed primarily by males of high genetic quality.
Women’s fertile sexuality, we have argued, is not merely heightened infertile sexuality. It is not merely the case that, at mid-cycle, women prefer more of all they prefer when they are infertile. Furthermore, contrary to Buller’s (2005) claim, fertile women are not simply more “sexual” than infertile women. Women’s preferences shift in a more fine-tuned way. And some preferences are heightened during the infertile phase, relative to the fertile phase.
As we discussed, women prefer male features valuable in long-term, stable mates just as much as or, in some instances, even more during infertile periods compared with fertile periods. Women’s preferences for kind, faithful men who would make good fathers do not shift or are more pronounced during the luteal phase. So too may be women’s preferences for men who possess genes that could diversify a family’s MHC alleles.
Furthermore, women do not possess uncontrolled and indiscriminate female proceptivity and receptivity during the fertile phase of their cycles. Rather, their libido is highly selective. As we argued in chapter 8, estrus in general has been misinterpreted by many biologists. In general across vertebrate species, estrous sexuality should be designed to lead females to be selective, except in response to rare circumstances.
The body of evidence we have reviewed is profoundly inconsistent with the common view that woman has no estrus. It is also inconsistent with the related view that woman’s sexuality has been liberated from the influences of ovarian-cycle hormones. That normally ovulating women experience cyclic shifts in sexual preference for putative markers of male genetic quality but users of hormone-based contraception do not establishes that reproductive hormones play a role in women’s estrus.
Tarín and Gómez-Piquer (2002) noted the similarity between estrous sexual motivation in female mammals in general and women’s mid-cycle sexual motivation implied by their high cycle-fertility preferences of the scent of men’s symmetry and masculinity. Their analysis is correct, we suggest, in claiming that women possess estrus. At the same time, it is incomplete because it does not recognize the functional significance of mammalian estrus, much less woman’s eroticism during their fertile phase. Woman’s estrous sexuality, specifically its proceptivity and receptivity, reflects preference for phenotypic markers of potential male genetic quality. We propose that this eroticism evolved because it led to the production of offspring of high genetic quality. Again, as we emphasized in chapter 8, women’s specialized sexuality during the fertile phase of the menstrual cycle is apparently homologous with and reflects adaptations functionally similar to adaptations of estrus in other mammals and female vertebrates in general.
Women, we argue, possess two functionally distinct forms of sexuality and associated adaptations for mate preference, specialized so as to be manifested at different times of the menstrual cycle (Gangestad & Thornhill, 1998; Little, Jones, Penton-Voak, Burt, & Perrett, 2002; Thornhill & Gangestad, 1999a; also see Penton-Voak et al., 1999; Penton-Voak & Perrett, 2001). The evidence we have reviewed indicates that one set of adaptations—which together constitute what we refer to as extended sexuality—functions at low- or zero-fertility menstrual-cycle phases. Its function is to increase access to nongenetic material benefits. The other set of adaptations—which together constitute what we refer to as estrous sexuality—functions during fertile phases to lead women to particularly prefer male indicators of genetic benefits for offspring. When pregnant or using hormonal contraception, women also manifest sexuality similar in design to that at infertile cycle phases. These two sets of sexual psychological adaptations of women show specialized function (Symons, 1987; Thornhill, 1990, 1997). To say that these two sets of adaptations have different functions is to say that each was ultimately forged by forms of past selection that were not causal in forging the other adaptation.
That is not to say that no adaptations overlap fertile and infertile phases. Given that women’s estrus functions in the context of a species in which females rely on material benefits delivered through stable pair bonds, women cannot completely ignore the costs they may pay for estrous sexuality in the currency of lost material benefits. Hence, though the forms of estrous sexuality we have discussed in this chapter have, we argue, been shaped largely by benefits in the form of good genes for offspring, women must also possess adaptations that operate even when women are fertile to avoid loss of a primary partner or his investment. In chapters 10 and 12, we discuss how women’s estrous sexuality operates in the context of pair-bonded relationships.
At the same time, when women do evaluate men as long-term partners, whether during fertile or infertile phases, they must concern themselves with characteristics that relate to both a flow of nongenetic material benefits that they can receive from a partner (including allocation of time to invest in a relationship or offspring) and genetic benefits for offspring. As expected, then, we see little evidence that adaptations for women’s long-term mate choices are expressed differently during estrus and during extended sexuality.
Before moving on, we discuss one final topic. It has been suggested to us that woman cannot possibly have female choice adaptation that functions only around ovulation, because ovulation is too uncommon an event to provide a context for effective selection for specialized ovulation-related adaptation. In natural-fertility human populations, such as the Dogon of Africa, the modal pattern of ovulatory cycles for a young woman is about two per 2 years (Strassmann, 1996a). In such populations, low-frequency ovulation in women stems from relatively frequent pregnancy, lengthy lactation, and associated infertility, as well as amenorrhea stemming from periodic paucity of nutrients and low levels of fat storage.
The effectiveness of selection in shaping adaptation, however, is not determined by the frequency of a problem that causes selection. Recurrence and significant selection pressure across generations are necessary for evolution by selection. Infrequent events coupled with large fitness consequences can generate major phenotypic evolution (e.g., Buss, 2005; Williams, 1992). (After all, individuals are born but once, but that surely does not imply that selection has not shaped adaptations specific to gestation and the birth process.) Sire choice based on genetic quality, we have argued, is an event with large fitness consequences. That the selection for preference in fertile-cycle-phase women for mates and sires of high genetic quality has been effective in leading to adaptations for such preference is indicated by the evidence, reviewed previously, that high-fertility-phase women have this form of special-purpose female choice. Said differently, fertile-phase women’s sexual preferences appear to show functional design for obtaining good genes for offspring and therefore the history of effective selection for such design. More generally, however, organisms commonly adapt to even rarer events than the infrequent ovulation in hunter-gatherer women. For example, the once-in-a-lifetime reproductive effort of semelparous plants and animals is characterized by special design shaped by selection. (In chapter 12, we address a related concern that ovulation by women not nursing offspring is a rare event after the birth of first child in traditional forager populations.)
Contrary to conventional wisdom, abundant evidence demonstrates that women possess estrus. In important respects, women’s estrus is functionally similar to, as well as homologous with, estrus is other mammals, and even vertebrates in general. The design of women’s estrus is to motivate sexual interest in men who are relatively masculine in scent, facial and bodily features, muscularity, voice, and skin tone; who have scent and appearance associated with bilateral symmetry; and who, perhaps, are relatively high in creative intelligence. These preferences are reasonably interpreted as sexual preferences for male features than connote good phenotypic and genetic quality (or did ancestrally, when the preferences evolved; see chapter 7). Hence the function of these adaptations of women’s estrus is to obtain through mating genes that enhance the reproductive value of offspring. Estrous women reveal additional specialized mate choice through aversion to maladaptive sexual behavior and avoidance of sexual coercion. (Aversion, avoidance, and rejection of males are expressions of female mate choice no less than mate preference is; e.g., Kokko et al., 2003.)
The pattern of mate choice demonstrated by women at the fertile phase of their cycles is not merely an exaggeration of preferences they possess during infertile phases of the cycle. During the infertile luteal phase, women may prefer relatively feminized men and men who appear healthy currently. The preferences of women who are pregnant or using hormonal contraception are similar to those of women in the luteal phase. All of these conditions are associated with high levels of progesterone. Women in infertile phases do not ignore indicators of good genes, as they are relevant to choice of a longterm, investing mate. Attractive women, who can garner long-term investment from a male of high quality, in particular value purported indicators of good genes when infertile. In general, however, preferences for long-term mates do not change much across the menstrual cycle.
