Sexual selection researchers refer to a variety of morphological features that capture the visual attention of conspecifics (notably, those of the other sex) but do not function as weapons or shields but as ornaments (Bradbury & Vehrencamp, 1998; the Latin root of the word means “to adorn or decorate”). Examples include patches of skin that are brightly colored, gatherings of colored feathers, feather plumes, and elongated or elaborated fins and tails. The same term, however, aptly applies to olfactory and acoustical traits that attract attention from conspecifics, notably in sexual contexts. In most animal species, males are the more ornamented sex (Andersson, 1994). Nonetheless, females in sex-role-reversed species, some socially monogamous species, some polygynous species with male territoriality or other mating systems, and some Old World primates possess ornaments. In Old World primates, female ornaments are referred to as sexual swellings—enlarged, often differentially colored anogenital areas. Old World primate females with multimale, multifemale social arrangements (e.g., most macaques, mandrills, baboons, and chimpanzees) typically possess sexual swellings, but so, to lesser degrees, do a monogamous gibbon and a harem polygynous langur (Dixson, 1998; cf. Pagel & Meade, 2006). Female sexual swellings occur in other mammalian taxa and at least one species of bird (see review in Dixson, 1998). Relatedly, some female primates and other species display a sexual skin, a differentially colored anogenital region without accompanying edema (swelling).
In primates, the color and size of ornaments typically peaks when females are near peak conception probability during estrus, though even pregnant females or otherwise infertile females (adolescent and anovulatory cycling females) may display ornaments (see Anderson & Bielert, 1994, and Dixson, 1998, for reviews; Deschner et al., 2004; Mohle, Heistermann, Dittami, Reinberg, & Hodges, 2005; Gesquiere, Wango, Alberts, & Altmann, 2007). Female sexual ornaments in primates and many other species generally are not displayed outside of a breeding context. (Throughout the remainder of the book, when we write of “conception probability,” we refer to probability of conception contingent on sexual intercourse—in humans, unprotected sexual intercourse.)
A wide range of findings on female ornaments exemplifies their important role in mating behavior. In the sex-role-reversed pipefish (in which females compete for male parental care), females develop a colorful venter during their brief breeding season. Males prefer more colorful females as mates, as color positively predicts female condition (Berglund, Rosenqvist, & Bernet, 1997; Berglund & Rosenqvist, 2001). Females of the lizard Sceloporus virgatus, a polygynous species, develop an orange throat patch during the breeding season. Males prefer more colorful females, which are in better phenotypic condition, have fewer parasites, and produce eggs of higher quantity and quality (Weiss, 2002, 2006). In the domestic chicken, another polygynous species, roosters prefer highly ornamented females, those with large combs. Though hens possess smaller combs than roosters, comb size in hens positively predicts condition, reproductive value, egg size, and yolk size (Pizzari, Cornwallis, Levlie, Jakobsson, & Birkhead, 2003). In the bluethroat (Amundsen, Forsgren, & Hansen, 1997) and barn owl, two socially monogamous bird species, large female ornaments are associated with superior female condition. In the owl, offspring of mothers with greater ornamentation inherit better immunocompetence, parasite resistance, and developmental stability than do offspring of less ornamented mothers (Roulin, 2004; Roulin, Jungi, Pfister, & Dijkstra, 2000). Emery and Whitten (2003) reported that swelling size of female common chimpanzees is positively correlated with ovarian function, measured as levels of ovarian steroids (most notably, estrogen). Finally, Domb and Pagel (2001) found that the size of sex swellings in female olive baboons positively predicts female condition, early age of maturity, number of offspring, and number of surviving offspring. Males, they also reported, prefer females with large swellings: Males particularly compete for, sometimes with costly aggression, and groom females with large ornaments. (See, however, other views and responses in Domb & Pagel, 2002; Nunn, van Schaik, & Zinner, 2001; Zinner, Alberts, Nunn, & Altmann, 2002; Zinner, Nunn, van Schaik, & Kappeler, 2004.) Male mate choice in savanna baboons (Hausfater, 1975; Packer, 1979; Smuts, 1985) favors sexual swellings of females that are maximally turgid and may be based, in part, on individual variation in swelling size (see Bercovitch, 2001). Moreover, laboratory experiments demonstrate that male baboons are more sexually motivated (as measured by masturbation frequency) by females with artificially exaggerated swellings than by females with normal swellings (Bielert & Anderson, 1985; Girolami & Bielert, 1987; also see Waitt, Gerald, Little, & Kraiselburd, 2006). (Other evidence of male mate choice in Old World nonhuman primates is summarized by Shahnoor & Jones, 2003; see also Alberts, Buchan, & Altmann, 2006; Gesquiere et al., 2007. Male mate choice in primates and other animals has not received as much attention from researchers as female choice. Yet it is known that mate choice by males can have dramatic positive effects on male reproductive success; see, e.g., Gowaty, Drickamer, & Schmid-Holmes, 2003, on house mice.)
These examples illustrate four generalizations about female ornaments. First, they are sexually selected as signals—they function to communicate information. Second, they are sexually selected through female-female competition for mates, male choice, or both. Third, they are rarely seen in species in which males transfer no material benefits to females; compared with other females, highly ornamented females often obtain more male material benefits. Fourth, the degree of ornamentation across females of a species often positively predicts condition or health in addition to number and survival of offspring. (For further evidence in a wide variety of species, see Amundsen, 2000; Amundsen & Forsgren, 2001; Amundsen & Pärn, 2006; Cuervo & Møller, 2000; Cuervo, Møller, & de Lope, 2003; Daunt, Monaghan, Wanless, & Harris, 2003; Griggio, Matessi, & Pilastro, 2003; Griggio, Valera, Cassas, & Pilastro, 2005; Hill, 2002; Jawor, Gray, Beall, & Breitwisch, 2004; Krebs, Hunte, & Green, 2004; LeBas, Hockman, & Ritchie, 2003; Siefferman & Hill, 2005; Swenson, 1997; Weiss, 2002, 2006.)
These generalizations are by no means hard and fast rules; some exceptions are notable. Ornaments of female red-winged blackbirds and barn swallows, for instance, may not have evolved through sexual selection (Cuervo, de Lope, & Møller, 1996; Muma & Weatherhead, 1991). In captive mandrills with food provisioning, evidence that female sexual swellings predict condition or are sexually selected is mixed (Setchell & Wickings, 2004, reported largely negative findings; Setchell et al., 2006, reported largely null findings in low-power analyses, but largely in a predicted direction; Dixson & Anderson, 2004, reported positive findings).
More generally, after a partial review of the literature on ornamentation across species, Cotton, Fowler, and Pomiankowski (2004) question whether special associations between degrees of ornamentation and individual health or condition have often been demonstrated. They point to the need for better experimental designs than have often been used to test these relationships. Though many studies reveal that ornaments hypothesized to function in sexual competition for mates are condition-dependent, few show that they are associated with condition more than nonornamental traits are (Cotton et al., 2004).
Though cognizant of the limitations imposed by the current literature and mindful of the need for more and better research, we defend the view that many, if not most, female ornaments have evolved as signals of quality or current reproductive condition. Though the view that ornaments function as honest signals of quality is widely represented and argued within behavioral ecology and the study of animals generally (see review by Searcy & Nowicki, 2005), the view is in fact not the traditional explanation of sexual swellings in Old World primates. Instead, female sexual swellings have traditionally been viewed as signals of fertility within the reproductive cycle. Much of the current chapter describes this view and argues that it is largely if not fully mistaken.
Darwin’s theory of female ornaments was that they are nonfunctional by-products of ornaments favored directly by sexual selection acting only on males (Darwin, 1871; see Andersson, 1994). That is, just as men purportedly have nipples because these features have been selected in women for specific functions and are nonfunctionally expressed in men, ornaments that have been selected in males may be nonfunctionally expressed in females. (We note here that we do not completely discount the possibility that nipples in men have been maintained because of their tactile sensitivity. Nonetheless, the idea that one sex may possess some features that are functional only in the other sex appears to be noncontroversial; see, e.g., Rice & Chippendale, 2001.) This theory may apply to instances in which female ornaments are muted forms of male ornaments (again, as male nipples are muted forms of female nipples). The model, however, cannot explain female ornaments in general (Cuervo & Møller, 2000; but see Cuervo et al., 1996; Muma & Weatherhead, 1991). Certainly, it cannot explain cases in which only females of a species exhibit an ornament—for instance, the anogenital sexual swellings in various Old World primates.
The conventional explanation of female primate sexual swellings, as well as the behaviors and odors produced by females in other taxa at fertile times of their reproductive cycle to which males sexually respond, is that these traits function to signal high fertility in the reproductive cycle (e.g., Aldridge & Duvall, 2002; Burt, 1992; Deschner et al., 2004; DeVleeschouwer et al., 2000; Hamilton, 1984; Hrdy, 1981; Jacobson, 1972; Liley & Stacey, 1983; Mason, 1992; Nelson, 1995; Scott & Vermierssen, 1994; Szalay & Costello, 1991). That is, females are presumed to be functionally designed to produce sex attractants, lay scent trails, or otherwise be conspicuous to assist males in finding them; otherwise, males might not be interested in mating, and females would not be inseminated. More generally, estrus in mammals and reptiles has often been similarly interpreted (on mammals, see Nelson, 2000, or general mammalogy textbooks; on reptiles, see Aldridge & Duvall, 2002, and the earlier references to this view cited therein). In chapter 8, we discuss and critique this view of estrus.
Women’s sexual ornaments (e.g., their enlarged breasts), of course, are not specific to the fertile phase of their cycles. The conventional interpretation of fertility signals nonetheless pertains to traditional explanations of women’s ornaments, as well, for the retention of women’s ornamentation across the cycle and their entire reproductive life spans has been interpreted by some scholars to mean that women have been selected to deceptively signal permanent cycle fertility (e.g., Szalay & Costello, 1991). Though, in some modified (but highly qualified) form, this idea may have some merit (see chapter 6), we argue that it is founded on a fundamentally incorrect view of female signals in general. Specifically, we argue that female sexual signals rarely evolve because they let males know when females are fertile. Put otherwise, female sexual signals rarely function to signal fertility in the reproductive cycle.
The appeal of the idea that females evolved to signal cycle-related fertility with sexual swellings is understandable (even if, as we argue, largely misguided). As we noted, in many species with such swellings, females do tend to exhibit them when fertile in their cycles. Though there is not, by any means, perfect correspondence between sexual swellings and cycle-based fertility status (as many primatologists have emphasized; e.g., Deschner, Heistermann, Hodges, & Boesch, 2003; Deschner et al., 2004; Engelhardt et al., 2005; Mohle et al., 2005; Reichert et al., 2002), positive covariation exists. Males who selectively seek copulations with females with sexual swellings therefore direct their efforts toward females capable of conceiving their offspring at better than chance levels. That is, males could have adaptive reason to pay attention to a signal correlated with fertility. Furthermore, the explanation proceeds, because females who are fertile need sperm to conceive, females could benefit by letting males know they are fertile. Hence females have been selected to produce a signal that males have been selected to attend to, as a means of gaining male attention (and sperm) just when they can use it.
Though appealing on the surface, on deeper reflection this view is almost certainly wrong, at least in this simple form.
Biologists widely recognize that, because of their different roles in reproduction, males and females typically have behavioral and physiological adaptations surrounding reproduction with different functions. Typically, females, as the sex that usually has a greater obligate investment in offspring (and, in mammalian species, always does), have adaptations that function to adaptively allocate their parental investment through control of that investment. By contrast, males typically have adaptations for fertilizing ova and thereby accessing parental investment held by the opposite sex (Alexander & Borgia, 1979; Bateman, 1948; Charnov, 1979, 1982; Eberhard, 1996; Emlen & Oring, 1977; Kodric-Brown & Brown, 1987; Queller, 1997; Parker, Baker, & Smith, 1972; Thornhill & Alcock, 1983; Trivers, 1972; Williams, 1966). This pattern holds true even in species in which males do invest substantially (though discriminately) in offspring, including in so-called sex-role-reversed species such as pipefish (Berglund, Widemo, & Rosenqvist, 2005) and tettigoniid katydids (Gwynne, 2001). Despite its nearly universal acceptance, this insight has implications that are often unappreciated or overlooked when it comes to interpreting female sexual behavior and ornamentation. The traditional interpretation that female primate sexual swellings and other female traits signal fertility is a prime case in point.
Parental effort is any effort allocated by an individual that increases the quality of offspring produced (Alexander & Borgia, 1979; Hirschfield & Tinkle, 1975; Low, 1978, 2001; Trivers, 1972). This effort obviously includes investment of time, energy, risk, and resources to support the growth and survival of offspring after they are born. But parental effort also includes investment of resources prior to birth, beginning with investment of resources into gametes. By definition, females produce larger gametes than males, and hence offer a larger initial investment in offspring. Though the initial asymmetry in female and male investment may disappear or reverse (see Gwynne, 2001; Trivers, 1972; Williams, 1966), it typically becomes exaggerated with investment beyond gamete production. In all eutherian mammalian species, internal fertilization and gestation require obligate parental effort on the part of females that far exceeds that of males. Indeed, in about 95% of mammalian species, males engage in very little or no parental effort (Clutton-Brock & Parker, 1992).
That is not to say that reproduction is not costly for males, as well, of course. In most species, however, males’ reproductive effort consists largely of mating effort, which is the male effort to find and attract mates and inseminate them, thereby increasing number of fertilizations. In most mammals, males expend far more mating effort than do females.
Two reasons that the sex difference in initial investment in parental effort becomes exaggerated are key to understanding most females and most males. First, as the sex that makes the greater parental investment, females have reason to be very discriminating about how they use that investment. They should be selected to make decisions that lead that investment to go far. If females mate with males who provide a poor complement of DNA, they have squandered that investment. Hence, females should be selective about whom they will mate with. When female mating is selective, the number of males that are eligible to mate (based on female preferences) is limited. Eligible males, then, can expect a relatively high future reproductive rate, which leads them to engage in mating effort—efforts to find and attract mates—rather than parental effort. Males who might benefit by parenting (because of a low expected future reproductive rate derived from mating effort) do not get the chance because females do not select them (Kokko & Jennions, 2002).
In some circumstances, biparental care has evolved. As we argued in chapter 4, humans appear to be one such species. In these species, females partly select males for their willingness to invest in parenting. Most or all mated males may engage in parental effort, and there is a smaller sex difference in allocation of effort toward mating and parenting. Nonetheless, some sex difference in parental effort typically remains. Males still have a smaller obligate investment in parenting with the production of each offspring. If they can attract females to mate with them without providing parental investment, they may benefit from engaging in mating effort in a way that females do not (e.g., Symons, 1979).
A second reason that the sex difference in initial investment typically becomes exaggerated is that the sexes often differ in parental certainty. The difference in parental certainty itself arises because of the asymmetry in initial investment. As females typically benefit from control over fertilization and additional investment into the offspring to an extent greater than males do, in many species fertilization takes place within a female’s reproductive tract (where she can, for instance, better exert control over who sires her offspring). In such instances, males are less certain about who their offspring are than are females. This asymmetry leads males to further devalue parental effort (which could be squandered on offspring not their own; Alexander & Borgia, 1979; Kokko & Jennions, 2002; Queller, 1997). They may engage in efforts to increase paternity certainty. This reason that parental investment becomes increasingly sexually asymmetric applies even when males do engage in substantial amounts of parental effort.
In sum, selection on females generally leads to adaptations designed to foster the efficient expenditure of parental effort. Females are hence designed to assess ecological circumstances that affect efficient expenditure of their parental investment. They are, furthermore, designed to differentially allocate parental investment depending on how circumstances affect its efficient use. These decisions should be sensitive to variations in season, resource availability, and mortality risks (e.g., diseases) that affect optimal allocation of parental effort. Females should also be designed to assess the quality of sires for their parental investments, for sire quality affects the profits females derive from their investments. When males do provide parental investment, females should be designed to assess and promote paternal investment into their offspring, which once again enhance the profits garnered from females’ own investments.
Selection on males, by contrast, typically leads to adaptations designed to secure the parental investment held by females. In many species, male reproductive success is maximized solely through ability to gain female parental investment; males succeed directly as a function of their ability to get females to invest in offspring they sire. Males in some species can also enhance fitness through their own parental efforts, but they get the opportunity to do so only if they are able to satisfy criteria by which females offering parental investment select sires; before they can engage in paternal investment, they must become fathers. (Obviously, it is also true that females cannot mother before they become mothers, but females typically need not satisfy stringent criteria of male choosiness simply to become mothers.) Hence males compete with other males to capture the parental investment offered by females. Naturally, selection on males should lead to phenotypic design for obtaining and assessing information about levels of parental investment females offer (whether it be condition-dependent, age-dependent, or cycle-dependent), including female attributes reflecting female phenotypic quality, and to allocating their mating efforts when those efforts are most likely to pay off. If they engage in parental effort at all, they should seek out and be sensitive to information pertaining to whether they have, indeed, successfully conceived an offspring (i.e., assess paternity certainty; see chapter 12).
As just noted, males should be designed to assess females’ ability and readiness to expend parental investment. Indeed, knowing when a female is about to expend parental investment is among the most valuable pieces of information a male can possibly possess. Sexual selection on males hence strongly favors abilities to perceive this information and act on it to secure female parental investment. These selection pressures are not recent; they have operated since the first females and males appeared in the history of life, that is, since the phylogenetic origin of anisogamy, and continuously thereafter (on the origin of anisogamy, see Parker et al., 1972; Kodric-Brown & Brown, 1987).
To appreciate, then, why females are unlikely to be selected to signal fertility in their reproductive cycles (at least with any signal that is costly to them), consider two females, one who signals and the other who does not. According to conventional thinking, the one who signals benefits because she lets males know when they should compete for her parental investment. Implicitly, this argument presumes that males do not know when the nonsignaling female is fertile (and furthermore that their not knowing hurts her reproductive success). A lone male who pursued copulations with the female who did not signal, however, could potentially outcompete males who ignored her, and, for that reason alone, great advantages to signaling fertility are not obvious. More fundamentally, however, if there are any incidental cues associated with the nonsignaling female’s fertility status (such as by-products of female fertility status), a male who perceived and acted on those cues would clearly be strongly advantaged over other males. Hence strong sexual selection on males should lead to adaptations to detect any cues of fertility status that exist (Thornhill, 1979; Thornhill & Alcock, 1983; West-Eberhard, 1984; Williams, 1992). We consider the evidence that such cues typically do exist and, therefore, that males typically know when females are fertile. (Instances in which selection favors female suppression of incidental cues—including, we argue, in humans—are exceptions. See chapter 11. Obviously these cases are not ones in which females purportedly signal fertility.)
Though females might be able to provide even better information to males through signaling, the marginal benefits that females can accrue through signaling may be slight (as males are generally already tuned in to incidental fertility cues) and very unlikely to offset the energetic costs of large signals. And that surely applies to most sexual swellings in primates, which are not particularly accurate cues of female ovulation (not obviously better than cues males have available anyway; e.g., Deschner et al., 2004; Engelhardt et al., 2005; Mohle et al., 2005; Reichert et al., 2002), yet costly. (We do not push the extreme argument that female swellings could not provide some incremental information about fertility status in the cycle to males and that females could not derive some small benefit from doing so. We do argue that the benefits of advertising fertility in the cycle per se are very unlikely to pay for the costs of sexual swellings. On the whole, other benefits most likely render those costs worthwhile.)
Based on precisely this line of reasoning, Pagel (1994) criticized the traditional view of sexual swellings of female primates as cycle-related fertility signals. In a quantitative game-theoretic model, he found that the sole evolutionarily stable strategy for males is to find and compete for fertile females—females that are at the point in their reproductive cycle at which conception can occur. Accordingly, females are not under selection to evolve costly traits that signal cycle-related fertility, such as sexual swellings.
If females emit incidental cues of fertility status, and if selection shapes male adaptations to detect and act on them, comparative data should yield much evidence for such adaptations (e.g., Thornhill, 1979). Females, by contrast, should rarely possess adaptations to detect reproductive readiness in males. In fact, these expectations are confirmed. Across a wide variety of species, males possess behavioral and related morphological adaptations to detect female fertile state and pursue females that are in that state. Females typically lack comparable adaptations.
Many of these adaptations are specialized sensors for olfactory detection, often at a great distance, of females that emit by-products of reproductive maturation. Male silkworm moths, for instance, have specialized antennae that can detect minute concentrations of a chemical that females release as a by-product during egg production (Jacobson, 1972). Bulls use chemicals in the urine of estrous cows, apparently breakdown products of the biochemical progression to peak fertility across the estrous cycle, to detect fertile females (Kumar, Archunan, Jeyaraman, & Narasimhan, 2000). Estrous female vervet monkeys excrete in urine prenanediol-3 alpha-glucuronide, a chemical related to ovarian function, which appears to be sexually attractive to conspecific males (see Szalay & Costello, 1991). In the long-tailed macaque, males’ sexual interest in females correlates strongly and positively with female estrogen levels and peaks at the most fertile phase of females’ menstrual cycles. Indeed, males prefer fertile-phase females having cycles that lead to conception over females in the comparable phase of nonconception cycles. Apparently, males perceive scent cues of estrogen levels (or their by-products) across the menstrual cycle (Engelhardt, Pfiefer, Heistermann, & Niemitz, 2004). In the chacma baboon, though males require several months of residency with a female to develop an ability to discern effectively her conceptive from her nonconceptive cycles, they ultimately do so, probably at least partly based on estrogen-related scent cues (Weingrill, Lycett, Barrett, Hill, & Henzi, 2003). Male chimps, too, can detect conceptive from nonconceptive cycles, presumably based on differences in estrogen levels (Deschner et al., 2004). So can male mandrills (Setchell, Charpentier, & Wickings, 2005), baboons (Alberts et al., 2006; Gesquiere et al., 2007), and stump-tailed macaques (Cerda-Molina, Hernandez-Lopez, Chavira, et al., 2006; Cerda-Molina, Hernandez-Lopez, Rojas-Maya, Murcia-Mejia, & Mondragon-Ceballos, 2006).
Functionally equivalent male adaptations that detect and track olfactory cues of female fertile states have been reported in algae, annelid worms, various arthropod taxa, salamanders, fish, lizards, snakes and multiple taxa of mammals (Aldridge & Duvall, 2002; Ferris et al., 2004; Liley & Stacey, 1983; Mason, 1992; Pickard, Holt, Green, Cano, & Abaigar, 2003; Preston, Stevenson, & Wilson, 2003; Rajanarayanan & Archunan, 2004; Roberts & Uetz, 2005; Scott & Vermierssen, 1994; Shine, Phillips, Waye, LeMaster, & Mason, 2003; Thornhill, 1979; Thornhill & Alcock, 1983; West-Eberhard, 1984; Williams, 1992). Behavioral adaptations to seek out this information are also widespread. Males of the vast majority of species of mammals, including most, if not all, Old World primates (excluding men in typical circumstances), exert considerable effort to touch, investigate, monitor, and smell the genital region of females (DeVleeschouwer et al., 2000; Dixson, 1998; Hoogland, 1995; Michael et al., 1976; Michael & Zumpe, 1982; Nelson, 1995; Takahashi, 1990; Weingrill et al., 2003). Once again, functionally similar adaptations to attend olfactorily, visually, or tactually to females’ genitalia or gonopores are found in males of nonmammalian vertebrates and even arthropods. (See Thornhill & Alcock, 1983, for a partial review of these male adaptations in insects.)
By-product cues must be distinguished from signals. Organisms respond to a great variety of stimuli or cues in their environments, only some of which are signals. A signal is a communicative adaptation—that is, a trait that has been selected directly because of its communication effect (Burghardt, 1970; Otte, 1974; also see Bradbury & Vehrencamp, 1998; Liley & Stacey, 1983). Put otherwise, a signal evolved because of information value it provides to other organisms (if sexual signals, members of the other sex), which those members act on in a way that benefits reproductively individuals who exhibit the signal. Communication—transmission of information to other organisms, which benefits the transmitter—is the function of a signal. Incidental or by-product cues can be emitted with no direct selection on emission of the cue for communicative value. These cues, then, are not signals. As Williams (1992) argued, chemical and other phenotypic stimuli emitted by fertile-phase females are rarely signals; instead, they are incidental by-products of changes to internal states of reproductive readiness. That is, these features are outcomes of physiological processes and structures that were shaped by selection for functions other than their communicative effects (e.g., in several cases discussed previously, structures designed to regulate estrogen levels as a means of affecting reproductive outcomes). Put otherwise, these processes and structures would have been favored by selection even if the incidental by-products detectable by males were completely absent. As we noted in chapter 2, to say that a feature is a by-product is to say that it had no role in bringing about the evolutionary persistence of the trait that gives rise to it as a by-product.
This argument does not imply that females do not benefit from by-products emitted incidentally at egg maturation or that correspond to high fertility in the reproductive cycle. They may, in fact, benefit through the attraction of suitable mates. But not every benefit of a trait is an evolutionary function, one that played a role in effectively shaping the trait (Symons, 1979, 1987; Thornhill, 1990, 1997; Williams, 1966, 1992; see also chapter 2, this volume). The benefit of attracting males typically played no role in bringing about the processes that led to by-products that attract males. In an evolutionary sequence, females emitted these by-products even before males used them as cues. Males then evolved specialized physiological, morphological, neural, and behavioral adaptations for finding and effectively inseminating fertile females on the basis of these by-products. The new benefits that accrued to females as a consequence of male adaptation typically had no effect on changes in frequencies of alleles underlying the development and operation of mechanisms regulating how those by-products were emitted and hence played little or no role in the evolution of these mechanisms. Females, in some sense, obtained these benefits for free, with no additional adaptation required. Hence females typically possess no specialized adaptation that functions to let males know they are reproductively ready.
Obviously, the claim that these stimuli are mere incidental effects is not to imply that they do not play important roles in male-female interaction and mating. They often play absolutely crucial roles. These roles, however, emerge through adaptation in males, not females (see also Williams, 1992). The fact that female emissions do play terribly important roles in mating unfortunately mislead many researchers into thinking that they must be signals. But they of course need not be and typically are not.
In the literature on nonhuman animal sexual behavior, writers frequently use the terms pheromone, signal, and advertisement to refer to stimuli emitted by sexually receptive females (for recent examples on primates, for instance, see Cerda-Molina, Hernandez-Lopez, Chavira, et al., 2006; Cerda-Molina, Hernandez-Lopez, Rojas-Maya, et al., 2006). All are inaccurate and misleading, as all imply that these stimuli function to attract males by communicating a message. In fact, these stimuli typically have no function in communication. Indeed, they have no function whatsoever.
Though writers typically imply that the communicative functional message of female stimuli that males track alerts males to reproductive readiness, other communicative functions are also claimed, most notably functions of communicating sex and species identity appropriate for reproduction (e.g., Dixson, 1998; Halpern, 1992; Mason, 1992). Though males may sometimes identify sex and species on the basis of scents and other fertility-related cues, these discriminations are by-products of selection for male abilities to detect females that are fertile and of high reproductive value. For instance, male garter snakes can sometimes discriminate the sex and species of other garter snakes based on olfactory cues (though evidence is mixed; see Mason, 1992). Males are highly attracted to scent of a female garter snake in her fertile phase, probably a by-product of the formation of mature eggs and associated estrogen (Halpern, 1992). They furthermore possess the ability to discriminate female body size and condition and prefer the scent of large fertile females in good condition (Shine et al., 2003). As males do not emit these same odors and females of different species have slightly different scents, males can discriminate sex and species based on the same cues. But these effects emerged as by-products of sexual selection on males to find and be attracted to fertile females (of their own species) in good condition. (For additional discussion of the reasons that sex and species identity are unlikely to be widespread functions of pair formation and courtship signaling systems, see Thornhill & Alcock, 1983; West-Eberhard, 1984.)
Our argument against female signaling of reproductive readiness has been largely conceptual. Females have no reason to expend large amounts of effort on signaling their readiness. Selection on males ensures that they will find fertile females based on by-products (or, at most, minimal efforts to disperse scents). Empirical evidence for female adaptation to signal, however, is also lacking. As we emphasized in chapter 2, an adaptation is a trait showing a specific functional organization/design because of a history of direct selection for that design. A directly selected trait—an adaptation—is one that solves a problem limiting individuals’ reproductive success and that, because of the solution it provides, is favored directly by selection. Demonstrating that a trait is an adaptation is an onerous task. It requires evidence for functional design in the trait to solve the problem. There is no compelling evidence, however, for functional design in typical vertebrate females to emit a vaginal scent at the fertile phase of the reproductive cycle to solve a problem that limits their reproductive success. Again, we have argued that females do not face a problem of getting access to limited sperm. Sperm for females is rarely limited (though selection, of course, does favor optimal allocation of sperm by males; see Wedell, Gage, & Parker, 2002). Furthermore, however, there is, for most species studied, no evidence of functional design to emit scents. When scents are signals, individuals typically have specialized machinery to manufacture and/or emit them, the chemicals produced typically involve complex molecules, and they are produced in relatively large quantities (see Bradbury & Vehrencamp, 1998). But, with few if any exceptions, females possess no specialized machinery to manufacture their vaginal scents. Their scents result from simple chemical-breakdown products of physiological changes associated with the onset of female reproductive status. They are emitted in small quantities. And females possess no specialized features to store, emit, or disperse scents.
Female signaling of fertility should be rare not only in animals but also in plants. Indeed, a classic case illustrates this point. The showy flowers of the angiosperms (flowering plants) do appear to be specialized structures that function to attract pollinators. A traditional interpretation of hermaphroditic flowers such as the rose or daylily was that they are functional analogs of female animal adaptations for signaling reproductive-cycle fertility, female adaptations designed to ensure fertilization of their ova. The interpretation shares those analogs’ theoretical problems. As Charnov (1979) first emphasized (see also Bateman, 1948), female reproductive success in plants is very unlikely to be limited by available pollen—by fertilization rate—because sexual selection on males ensures high rates of fertilization of female gametes. Rarely will females in flowering plants experience a net benefit from developing costly structures that function to attract pollinators. Showy flowers, then, are typically sexually selected male adaptations that increase male “mating success” by attracting pollinators (Andersson, 1994; Bell, 1985; Charnov, 1982; Willson & Burley, 1983).
Knight et al. (2005) thoroughly reviewed effects of pollen limitation on female reproduction. Many cases have been documented. Knight et al. (2005) suggest, however, that these instances may typically arise from evolutionarily novel circumstances. For instance, introduction of a plant into a new ecology (a common practice) may also introduce novel distributions of and densities of plants and pollinators, which then give rise to pollen limitation on female reproduction. Pollen limitation in novel circumstances, however, by no means implies that the same plants evolved under circumstances of pollen limitation.
Indeed, a classic purported case of sperm limitation has recently been shown to be an outcome of evolutionary novel conditions. Reproduction of the roundworm, Caenorhabditis elegans, a commonly studied species, is sperm-limited in laboratory conditions. Females commonly lay many unfertilized eggs; males do not deliver adequate sperm numbers to fertilize all eggs. As the title of their article announces, Goranson, Ebersole, and Brault (2005) have resolved the “adaptive conundrum” this phenomenon poses. In favorable lab conditions, females achieve unusually large body size, which leads them to lay very large numbers of eggs. In harsher natural conditions, females grow to smaller size, they lay fewer eggs, and available males fertilize them. Females’ reproductive success is simply not sperm-limited in natural conditions.
Similarly, evolutionary novelty may explain examples of sperm limitation in broadcast spawning invertebrates and algae (Yund, 2000). In these instances, mobile male gametes seek out female gametes, typically using chemical cues. Industrial pollutants may kill or disrupt adaptive chemotaxic orientation of male gametes (e.g., see Meric et al., 2005).
Male gametes are ardent and, similar to males themselves seeking females, sperm often use chemical cues to find eggs for fertilization in the female reproductive tract. Polyspermy (entry of multiple sperm into an egg at conception) is widespread in organisms and occurs in humans. It is maladaptive for both sexes, as it leads to death of the zygote or maladaptive development. Polyspermy can be understood as a maladaptive by-product of male adaptation to conceive before sperm (1) die, (2) are outcompeted by sperm of another male, or (3) are rejected by the ovum or female reproductive tract. Selection on females of various taxa has crafted counteradaptations that function to prevent polyspermy (e.g., defenses against multiple sperm entering the egg, reduction of sperm numbers in the reproductive tract; for evidence of adaptation of eggs against polyspermy in humans, see Patrat et al., 2006). These counteradaptations may well have maladaptive by-products themselves: infertility. As Arnqvist and Rowe (2005) note, the line between a female’s effective prevention of polyspermy and her infertility due to overprotection of eggs against polyspermy can be a thin one. High rates of female infertility in species (not atypically 10–15%; Arnqvist & Rowe, 2005) is not due to insufficient quantities of sperm; it is more likely due to female adaptation in response to the overabundance of sperm, which functions to prevent polyspermy. (See also Gomendio et al., 2006.)
If, in fact, females rarely signal cycle-related fertility, the many instances in which researchers have explained phenomena as outcomes of female adaptation to signal cycle-related fertility to males typically reflect misinterpretation. Two kinds of misinterpretation are common. The first kind is one that sees female adaptation to signal when no such adaptation exists. Again, males are adapted to respond sexually to incidental-effect cues of female fertility. The second kind of misinterpretation is one that correctly recognizes female adaptation to signal but misinterprets the function of the signal, viewing it as a signal of fertility in the reproductive cycle when, in fact, it is a signal designed to honestly communicate information about quality.
Female Scent as a By-Product As already noted, researchers commonly interpret female by-products to which males are attracted as signals, when in fact they have no function. Aldridge and Duvall (2002), for instance, interpret the function of the scent of fertile-phase female rattlesnakes derived from egg maturation as “advertising female receptivity” and “assisting males in locating” females. Elsewhere in the same article, however, these authors document that males are elegantly designed to find estrous females (also see Duvall & Schuett, 1997).
In many species of insects and the Asian elephant, females at peak fertility produce and release a particular molecule to which males are attracted. The chemical has been hypothesized not only to signal peak female fertility but also to function to sexually motivate males and synchronize their mating behavior with that of females (Rasmussen, 1999). The widespread occurrence of the chemical across taxonomically distant females has been interpreted as evolutionary convergence. That is, the same selection pressure purportedly acted independently to create the same female adaptation involving this chemical in distantly related species (Rasmussen, 1999; Rasmussen, Lee, Roelofs, Zang, & Daves, 1996). Moreover, the compound’s volatility has been argued to be a key reason that it serves the function of communicating peak fertility well and, as a result, has been selected to do so in different taxa (Rasmussen et al., 1996).
A simple by-product hypothesis accounts for the widespread distribution of this chemical just as effectively. This chemical may simply be a common by-product of physiological achievement of peak female fertility, which males have evolved to detect and be attracted to. As predicted by this hypothesis, the chemical has a simple structure, and females possess no specialized apparatus to emit it. Indeed, in the Asian elephant it appears to be “manufactured” in the liver (Rasmussen, 2001). The liver, of course, is where many by-products of metabolic processes are broken down, released into the bloodstream, and excreted in urine. Males do, in fact, detect this chemical in urine; it cannot be detected in the female vaginal area. The liver need not possess a special adaptation to create this chemical so that males can detect it in urine. It merely needs to do its jobs, which involve preparing waste products for elimination. Rasmussen et al. (1996) are probably right when they suggest that the chemical’s volatility is a reason that males use it as a cue to discriminate the fertility status of females, but for the wrong reason. Instead of females having been shaped to produce a “signal” that, due to its volatility, tracks female fertility, males evolved to detect and be sexually motivated by a chemical that, due to its volatility, affords effective tracking of female fertile states.
In the same species of elephant, bulls emit a diverse array of specialized chemicals that regulate sexual competition among males and are associated with males’ relative status in the hierarchy of local males. The traits responsible for production and dispersal of these chemicals do reveal evidence of functional design for communication as honest signals of male quality (Rasmussen, 1999; Rasmussen, Krishnamurthy, & Sukamar, 2005; Schulte & Rasmussen, 1999). The chemically simple sex attractant of female elephants, however, does not provide evidence of any function.
Recently, Rasmussen and colleagues have identified a purported “pheromone” in the urine of female African elephants, one that is shared with a species of beetle (Goodwin et al., 2006). It too permits males to detect a female’s preovulatory phase (Bagley, Goodwin, Rasmussen, & Schulte, 2006). Once again, this chemical is probably simply a by-product of a metabolic process. It is not a signal or a pheromone, and its production has no function.
Female Ornaments as Signals of Quality In some species, females do appear to produce chemicals that function as signals. In these instances, however, these chemical signals are unlikely to have evolved because they signal peak female fertility. Rather, their function is probably to signal individual female quality. In tamarin monkeys, both sexes parentally invest. Because males invest in offspring, females compete for male parental care. Males have accordingly evolved to prefer to mate with and invest in the offspring of high-quality females. Put otherwise, in this mating system, mutual mate choice has evolved due to selection on the preferences of both sexes (e.g., Kokko & Johnstone, 2002). Heymann (2003) argues that sexual selection on females in tamarin and related New World monkeys has led to specialized scent glands and scent marking behavior that function to honestly signal quality. (At the same time, full evaluation of these claims demands additional research. See Dixson, 1998, for a review of literature on scent glands and scent marking in female primates.)
Female isabella moths produce a chemical sex attractant. In contrast to most other female moths, however, females in this species produce copious quantities of the chemical in unusually elaborate glands, which are then released in visibly aerosol form. In contrast to most male moths, which respond to a few molecules of sex attractant, male isabella moths sexually respond only to large amounts of the scent. Furthermore, females in this species actively court males more than males court females, atypical of moths. Indeed, the courtship pheromonal system used by males, elaborately designed to communicate with potential mates in most moths, is vestigial in the isabella moth.
The mating system of this moth appears to involve sex-role reversal with mutual sexual selection, but greater sexual selection on females than on males (that is, more selective mate choice by males than by females; see Gwynne, 2001, for a discussion of sex-role reversals in insects). As should be expected if females are sexually selected to attract males, males invest substantially in offspring. Specifically, they offer nuptial gifts of pyrrolizidine alkaloids, toxins that function to provide aposematic predator defense. As males may collect and provide differing amounts of alkaloid, females may compete for males who provide much. Indeed, in a closely related species, females prefer males with relatively large loads of the alkaloid. Males able to deliver large loads of alkaloids can therefore afford to be more choosy than typical male moths, which, we suggest, has led to adaptations underlying male mate choice and female competitive signaling of quality, including through costly production of an aerosol sex attractant (see Krasnoff & Roelofs, 1988, 1990; Krasnoff & Yager, 1988.) Consistent with this interpretation of sex-role reversal and female competitive signaling of quality, Lim and Greenfield (2006) recently reported the remarkable discovery that displaying females in this moth form leks.
Honest-signal female ornamentation in the form of pheromonal communication probably occurs in other moth taxa. Honest signaling should be seriously entertained whenever females possess complex structures to emit or disperse scents, as seen, for instance, in certain female arctiid and noctuid moths. We note, however, that mere presence of female glands that produce or store sex attractants is not evidence sufficient to conclude that females possess ornaments. Glands and their secretions function in numerous contexts. Glands that function to eliminate waste, for example, may excrete products that attract males if those products are produced primarily or only by fertile females. In such cases, no adaptation for female signaling exists. Indeed, females of the leafminer moth, a parthenogenetic species (and hence totally lacking males), engage in scent-dispersing behavior, possibly due to excretion (Mozuraitis, Buda, Liblikas, Unelius, & Borg-Karlson, 2002). Obviously, the scent does not attract males, as none exist. In a phylogenetically closely related two-sex species, however, it does. Despite the evolution of parthenogenesis in the leafminer moth, scent-dispersing behavior was retained, probably because it functions in a context other than mate attraction.
We have emphasized that female signaling of fertility within the reproductive cycle should be rare in animals, at least in any conspicuous form. Under very limited circumstances, however, it may evolve. Specifically, if males are rarely encountered and highly dispersed, females who signal fertility and thereby reduce time to search for a mate could be favored by selection (e.g., when populations are small, individuals are widely dispersed, and the adult sex ratio is female-biased due to high male mortality). Linyphia litigosa spiders may be an example. Females in this species release a chemical that may function as a pheromone that signals fertility in the reproductive cycle to males. Interestingly, they have been observed to release pheromones only under unusual circumstances—when they are unmated, full of eggs, and isolated experimentally from males for several days, situations that may simulate the natural atypical circumstances of small, marginal populations late in the year after most males are dead (Paul Watson, personal communication, October 15, 2004; 1986). The possibility that the putative sex pheromone of female Linyphia is a by-product scent produced as a result of physiological changes associated with the absence of mating and/or with advanced egg age inside the female has not yet been ruled out. The alternative hypothesis, that Linyphia females engage in competitive signaling of individual quality, seems unlikely, as males have little opportunity to compare females and thus choose mates under the circumstances in which the scent is released (specifically, very low population density).
We have emphasized that, despite rare exceptions, females do not commonly possess costly adaptation that functions to signal their fertility in the reproductive cycle. Similarly, we anticipate that females will lack costly adaptation that functions to signal nonreceptivity. A signal of nonreceptivity in the cycle is equivalent to a signal of receptivity (in that, in the former case, males know that a female can conceive when not signaling). This is not to say that females lack adaptation to resist males when infertile in their cycles, as well as when fertile. They do, as documented across animal taxa (Arnqvist & Rowe, 2005). For example, nonreceptive female tenebrionid beetles spray a noxious, disabling spray on persistent suitors. Sprayed males fall into a “coma” for several hours, which permits females to escape. The spray and spraying behavior are adaptations, though primarily used against predators. Female houseflies have a spine on their middle legs that they jab into the wings of unwanted mates. Overly persistent males may pay for their rudeness by having their wings shredded. This female adaptation may function to serve mate choice and/or to reduce costs of insemination (see Thornhill & Alcock, 1983, for these and other examples in insects). Female alpacas spit, kick, and bite males that try mating outside peak estrus. Both sexes engage in the same agonistic behaviors toward other llamas and even toward humans who bother them (Vaughn, Macmillan, Anderson, & D’Occhio, 2003). Female domestic dogs and cats are notorious for their outright hostility toward ardent males that attempt intro-mission outside peak estrus. (See Arnqvist & Rowe, 2005, for a review of resistance and rejection by females in the context of sexually antagonistic interactions.)
These adaptations function not as signals of nonreceptivity but instead to reduce harassment by ardent males and other unwanted social interaction, to lower risk of predation and costs of insemination when infertile (see Arqvist & Rowe, 2005; also see chapter 7, this volume), and, when fertile, to avoid insemination by unsuitable males. Resistance or rejection is not a “signal” to males of nonreceptivity or infertility, as some biologists have conjectured (see Ringo, 1996; Schuett & Duvall, 1996). That is to say, its costs are not (largely) paid for by information transmitted to males. Instead, its primary benefit is the benefit of rejection of males itself.
Females, we have claimed, rarely if ever have adaptation that functions to signal fertility in the reproductive cycle to thereby ensure insemination and conception. Sexual selection on males ensures insemination and conception of fertile-phase females. An alternate but related form of the hypothesis is that females signal fertility not to obtain any insemination but, rather, to ensure that males with the best genes inseminate them. The basic idea is that fertility signals function to incite male-male competition. Therefore, males with the best genes compete to inseminate a signaling female. Though a female could surely be inseminated without signaling, this argument acknowledges, she improves the quality of her offspring by inciting competition. This idea was, in fact, first introduced as an explanation of female sexual swellings in primates (Clutton-Brock & Harvey, 1976).
This hypothesis encounters precisely the same problem that any fertility signal hypothesis has: It assumes that males do not know which females are fertile without an overt, costly signal. Just as sexual selection on males ensures that males will detect extant by-product cues of female fertility status as it changes across the cycle, so it ensures competition between males to inseminate females when they are fertile. Males do not typically require a special signal of fertility to be interested in competing for copulation with fertile females.
Indeed, according to this hypothesis, dominant males should not dominate matings with fertile females in primate species lacking female sexual swellings, as they lack a key sign of fertility to motivate them. In fact, precisely the opposite is found: Dominant males overrepresentatively mate with females at peak fertility even in nonhuman primates that lack sexual swellings and skins (Baker, Dietz, & Kleiman, 1993; see a review of other studies in Dixson, 1998). This same pattern exists in nonprimate mammals in general, despite the absence of sex skins and swellings in the great majority of these species (e.g., Ginsberg & Huck, 1989; Preston et al., 2003).
As Pagel (1994) keenly pointed out, this idea also faces a problem explaining the costliness of a purported fertility signal. Suppose males did respond to and compete for females displaying a costly fertility signal. Females who possessed less exaggerated forms of the signal would presumably lose out. Imagine now, however, a dominant male who does not compete for females with costly signals but rather copulates with females that exhibit weaker cues. If, in fact, the cue tells nothing more than fertility status (i.e., it does not relate to female quality or reproductive condition), this male wins out, for he has paid minimal costs for competition relative to other successful males. And these females win out, for they obtained a dominant sire but paid fewer costs for their signals. Mere incitation of competition cannot sustain costly fertility signals. Put otherwise, costly signaling of fertility status to incite male interest is not an evolutionarily stable strategy. Fertility cues should hence typically be cheap—and, indeed, as we have just seen, in many species females pay no special costs for them whatsoever; they are mere by-products of adaptations with different functions.
Other purported fertility cues in nonprimate species have also been mistakenly interpreted as fertility signals designed to incite competition. Females of some avian species (such as alpine accentors) sing only when fertile, thereby attracting males (see Montgomerie & Thornhill, 1989). Similarly, female bearded tits exhibit “chase-flights” during the fertile phase to attract males. Though often interpreted as fertility signals (e.g., Cox & Le Boeuf, 1977; Hoi, 1997; Langmore, Davies, Hatchwell, & Hartley, 1996), these signals, we propose, may be ornamental signals of individual female quality. Indeed, chase-flight behavior by bearded tit females is energetically demanding. As expected if it is a quality signal, females in good condition engage in it more than do females in poor condition (see “Female Ornaments as Honest Signals of Quality,” below).
Consider another female behavior claimed to function to incite male-male competition: homosexual mounting behavior of estrous female cattle and estrous females in certain other mammals (Adkins-Regan, 2005; Parker & Pearson, 1976), including nonhuman primates (O’Neill, Fedigan, & Zeigler, 2004). Parker and Pearson (1976) argued that, by mounting other females, females let males know they are in estrous and thereby garner bulls’ attention. Bulls are indeed attracted to female homosexual mounting of females (Geary & Reeves, 1992). Once again, however, this behavior is better interpreted as an ornament signaling quality by displaying the ability of a female to dominate another female. In fact, only some estrous females exhibit the behavior (VanVliet & VanEerdenburg, 1996).
One of the traits first offered as a female adaptation for inciting male-male competition was described by Cox and Le Boeuf (1977). Female northern elephant seals protest the vast majority of copulations males attempt. They are particularly likely to protest copulation attempts by subordinate males. When a female protests, males dominant to the offending male often intervene (at times, competing with each other to do so). As Cox and Le Boeuf note, the effect of the protest is indeed to perturb male behavior. Accordingly, they suggest that the function of a protest is that “it literally wakes up sleeping males and prompts them to live up to their social obligations” (1977, p. 328). As a result of interventions initiated by protests, females end up with sires that offer better genes for offspring.
As Cox and Le Boeuf also note, however, if females did not protest, “males would still compete and interfere with each other’s copulations” (1977, p. 328). And, indeed, it makes little sense to think that males are, in a sense, asleep on the job; surely, males who were not so lax would be selected over “sleeping” males. It makes more sense to think that, in fact, males are highly vigilant to ongoings but that, in large aggregations in which a few males can garner most matings, they must simultaneously monitor many events. The fact that they do respond to protested copulations with intervention suggests that males do indeed monitor their surroundings for key fitness-relevant events. A protested copulation brings attention to an event that calls for an adaptive response (intervention)—but it does not initiate a motive to compete with other males in an otherwise stupefied male.
Female protests in this species are not signals of fertile states. Indeed, females protest virtually every copulation attempt prior to their fertile phase. As they near ovulation, they selectively do not protest mounts by dominant males. Their protests have benefits of reducing the direct material costs of copulation (during which females can be injured; see Cunningham, 2003) and increasing genetic quality of offspring (partly through the mere effect of rejecting a low-quality male, partly through intervention by self-interested dominant males). It does not do so, however, by waking up sleeping males.
In some species of Old World monkeys and apes, females vocalize during or soon after mating (Dixson, 1998; Maestripieri & Roney, 2005; Pradhan, Englehardt, van Schaik, & Maestripieri, 2006). The prevalence of female calls varies greatly across species and studies, ranging from 6 to 99% of all copulations. Some researchers have claimed that female calls function to incite male-male competition (see Maestripieri & Roney, 2005, for a discussion). Maestripieri and Roney (2005) specifically proposed that these calls function to promote postcopulatory mate guarding by preferred males and thereby reduce the likelihood of mating with nonpreferred males when females are fertile in their cycles. Indeed, they claimed that copulation calls themselves are honest signals of fertility. Consistent with this hypothesis, the occurrence and intensity of female calls are associated with high levels of sexual swelling (i.e., at peak estrus and fertility), matings with dominant rather than young or subordinate males, and the timing of male ejaculation. Males, furthermore, are most likely to respond to calls when females are in estrus. In Barbary macaques, males are more attracted to recorded calls of females at peak fertility than calls emitted outside peak fertility in the cycle (Semple & McComb, 2000).
Pradhan and colleagues (2006) offer related but distinct interpretations. Copulation calls function by alerting other members of a primate group—notably males—that a female has copulated. (Naturally, many will have attended to the act, but some may have been attending to other females, feeding, or engaged in other activities.) Copulation calls do typically lead other males to approach a calling female and, should she not be attended by her consort, reduce the amount of time passed before she mates again. A copulation call, hence, can, in effect, increase the benefit of a dominant male to attend to a female he has just inseminated, which can increase his confidence of paternity in the offspring and provide her subsequent material benefits (e.g., his protection of the offspring), as well as reduce her cost of resisting inferior males. Naturally, this strategy will work best if the male suspects that the calling female is fertile. But by no means need that imply that the copulation call itself is a fertility signal. Males can detect fertility cues. If a copulation call occurs in the context of the calling female emitting such cues, the payoffs to a male guarding her postcopulation are greater than if she emits no fertility cues. Hence, the strategy of calling is more likely to pay off when a female is fertile. Co-occurrence of calling with peak fertility is not an indication that calling signals fertility. Indeed, as Pradhan et al. (2006) stress, females may call when infertile or pregnant, as well.
Pradhan et al. (2006) speculate that, in species in which females call after nearly every copulation (not just those with dominant males), females may benefit by increasing sperm competition. In such cases, dominant males do not tend to guard after a female calls and she mates again more quickly. The argument is not that females are sperm-limited (or even limited by good DNA) but, rather, that females may benefit by running sperm races or selecting compatible genes in utero (e.g., Zeh & Zeh, 2001; see also chapter 7, this volume). Just as female elephant seals call other males’ attention to fitness-relevant events by protesting copulations, these female primates may call males’ attention to fitness-relevant events to which they may adaptively respond. No assumption that females “wake up” sleeping males, however, need be made. Whether adaptive in utero selection for sperm occurs in primates is unknown, and, hence, this hypothesis requires additional evaluation. In any case, however, in these species copulation calls occur throughout females’ receptive periods (both fertile and infertile phases) and hence clearly do not function as fertility signals.
As we discuss later, female ornaments possibly do influence male interest in competing for females and thereby affect the quality of genes of her offspring’s sire. They do so, however, not because they signal fertility. Rather, under special circumstances, they could influence male choices about how much and what kind of mating effort to exert to mate with their female bearers because they signal female quality.
We do not rule out the possibility that low-cost signals that garner attention from males when females are fertile could potentially evolve, particularly when males are widely distributed spatially. Some female mammals vocalize during estrus. In one well-studied example, female African elephants emit the “estrous call” at peak fertility and before mating. Dominant, but not subordinate, males are attracted by the call (Poole, 1999). These calls may be low cost to females; by no means can one safely generalize this example to instances of costly signals such as primate sexual swellings. The alternative hypothesis that these calls communicate information about female quality should also be evaluated.
Finally, to say that females rarely signal cycle-related fertility to incite competition between males to garner sperm or good genes is not to say that females lack adaptations that function to efficiently bring about reliable contact with and conception by males of superior genetic quality while reducing contact and mating with inferior males. Such adaptations are straightforwardly expected by basic understandings of female mate choice. Female resistance against certain ardent males, as discussed earlier, is one kind of such adaptation. In addition, fertile-phase females may possess adaptations that affect their habitat preferences and movement patterns to increase contact with males of high genetic quality and simultaneously reduce or eliminate contact with low-quality males. Female guppies, for example, select microhabitats that ensure reduced contact with low-quality males and enhanced contact with high-genetic-quality males when in their fertile phase (Kodric-Brown & Nicoletto, 2005). The microhabitats preferred by fertile females are more costly for males, especially low-quality ones, than for females to inhabit. But these microhabitat preferences do not function to incite male-male competition. Male guppies compete for fertile-phase females in any natural habitat and are highly sexually motivated by a by-product scent emitted by these females (see chapter 8 for additional discussion). Rather, these habitat preferences affect females’ relative rates of encounter with males of high and low quality.
We began this chapter by noting that, across a wide variety of species, many traits deemed attractive and the perceptual adaptations that deem them so appear to have evolved as components of honest signaling of quality. Quality can refer to either phenotypic traits or genotypic features. In the context of mate choice, phenotypic quality refers to features that enhance the value of an individual as a mate (e.g., ability or willingness to provide current investment, as affected by current condition or indicators of future condition; condition-based fertility; reproductive value; ability to provide genetic benefits to offspring). Genetic quality specifically refers to ability to deliver genetic benefits to offspring. The differential cost to bearers of honest signals and/or the differential utility of signals across the life span renders the degree of a signal elaboration an honest or reliable and stable display of quality (see review by Searcy & Nowicki, 2005; also Getty, 2002, 2006; Grafen, 1990; Kodric-Brown & Brown, 1984; Zahavi & Zahavi, 1997). Sexual signals often are costly in terms of energy, predation risk, or the fact that they involve or invite social competition from same-sex others. Though much more attention has been directed to understand male signals of quality (for the reason that they are much more common than female signals), signals of quality may evolve in females as well. If how much females pay for or get out of signals of a certain size (in currencies of fitness gains) depends on their condition, how “big” (large, intense, or otherwise costly) a signal it is worth females to pay for will vary systematically with their condition; for those in better condition, it will pay to build a bigger signal (e.g., Getty, 2002, 2006; Grafen, 1990). Of course, costly signaling cannot evolve without perceivers’ responses to pay for signals’ costs. And, naturally, perceivers’ responses will not be directly selected unless they promote perceivers’ reproductive success. As males evolve to be attracted to females with bigger signals (because they covary with female condition), the signaling system becomes stably reliable. At that point, females possess adaptations to signal as a function of condition. And males have adaptations guiding their mate choice based on female signals’ degree of elaboration.
How reliable signaling systems become stabilized once signals discriminate individuals’ quality is not difficult to appreciate: Once that condition is met, it pays mate choosers to choose on the basis of the signal, which reinforces the signaling sex to possess a big signal and stabilizing the signal as a reliable indicator of quality. Harder to explain is how the signal becomes predictive of quality. It cannot simply be that mate choosers prefer the big signal; if they did, it would be an indicator of quality. Mate choosers, after all, cannot evolve to prefer the big signal as a quality indicator until it has already become predictive of quality. Obviously, mate choosers cannot know ahead of time that the signal will become a signal of quality if only they would prefer it sufficiently (and, even more important, will not be benefited for doing so until it is a signal correlated with quality). For a trait to evolve as a signal of quality, it must somehow be correlated with quality before it actually qualifies as a signal of quality (a trait that evolved because mate choosers preferred it as a quality signal). Sexual selection theorists propose two main scenarios by which traits become associated with quality and thereby become signals of quality.
The first is the route of preferred signal through sensory bias. In this scenario, a trait first becomes preferred not because of any fitness advantage to the mate chooser for preferring it. Instead, the preference for the trait is merely a by-product of a sensory adaptation of the mate chooser that has a function unrelated to mate choice. For instance, perhaps mate choosers are drawn to red objects because ripe fruits are red (i.e., “attend to red” reflects a food preference adaptation). Potential mates that exhibit some redness, though not edible, get attended to, which gives those potential mates some advantage on the mating market. As the preferred trait becomes exaggerated through mating advantages of those who have it (e.g., as signals become increasingly red or large), however, individuals in best condition are able to effectively display it in its most preferred form. At that point, the trait has become an indicator of quality, and preferences for the trait no longer need to be maintained as by-products of nonmating adaptations (see Kokko, Brooks, Jennions, & Morley, 2003; for a recent discussion of sensory bias and other sexual selection models, see Fuller, Houle, & Travis, 2005).
The second route is that the preferred trait is associated with quality prior to its evolving as a signal. Individuals in better condition generally have greater amounts of energy to allocate to traits important to survival and reproduction. Hence, a wide variety of traits may actually discriminate individuals of different quality. In many species, however, as the amount of available energy increases, a larger proportion is allocated to traits that foster immediate reproduction. Individuals have but one life to give. When circumstances threaten survival, it often makes adaptive sense for individuals to protect that one life by engaging in mortality reduction efforts (e.g., sequestering energy reserves to maintain survival). When individuals’ condition is more favorable, they may reproductively benefit from putting a greater amount of energy into reproductive traits. That certainly happens in human females, whose estrogen levels and fertility levels are very sensitive to their energy stores and energy balance (e.g., Ellison, 2001). Only when their caloric intake consistently exceeds demands of energy for maintenance of survival functions do estrogen levels increase sufficiently to render successful conception and implantation likely. Certain traits—often reproductive ones—thus may tend to particularly differentiate individuals varying in condition, simply due to the way that individuals have been shaped to optimally allocate energy to their different traits. At this point, the traits that covary with condition do not do so because they have evolved through sexual selection as signals. Because they do covary with condition, however, selection can lead to adaptations in mate choosers to attend to and prefer mates who exhibit these traits. When such adaptations evolve, the quality-discriminating traits are sexually selected for their signal value, as well as selected for functions they had prior to being valued by mate choosers. Their added benefit as signals should lead their bearers to allocate more effort into developing them than they would otherwise. These traits thereby become exaggerated as displays of condition through sexual selection (see also Rowe & Houle, 1996).
To say that a signaling system is reliable or honest is not to imply a complete absence of deceptive signaling as a component of the system. Signaling systems that have been investigated in detail reveal largely honest signaling with an admixture of deceit (see review in Searcy & Nowicki, 2005). Signaling systems require honesty to be stable because only honest signals lead selection to maintain perceivers’ responses, which in turn maintain benefits to signalers. In light of the complexities of some signaling systems, some signalers can gain advantages by signaling dishonestly (i.e., the magnitude of their signals does not correspond with their quality). As well, some signals may be errors: An individual may produce too small or large a signal for its current quality, but not for an earlier ontogenetic state when the signaling trait was developed. Furthermore, signals may be honest in some contexts but not others (e.g., a signal may be an indicator of quality in younger but not older individuals; Kokko, 1997). But even in the presence of deceptive signaling, perceivers’ responses to signals must promote their own fitness, on average, for the signaling system to persist (e.g., Kokko, 1997). Put otherwise, dishonest signaling in animal communication can only persist in the context of a largely reliable signaling system (Searcy & Nowicki, 2005).
In chapter 7, we discuss in greater depth the literature examining whether male sexually selected traits are indicators of condition. At that time, we also delve into issues pertaining to whether these indicators of condition benefit choosers through good genes or direct benefits, as well as variations on the evolution of honest signals of good genes (e.g., the Fisherian process).
If female sexual swellings are not signals of cycle-based fertility or instruments designed to incite competition between males by advertising fertility, how did they evolve? Pagel (1994; also Bercovitch, 2001) hypothesized that the swellings are adaptations that function to display female phenotypic and genetic quality. Males that preferred the exaggerated forms of display experienced a net reproductive benefit, and hence the male preference for exaggerated displays evolved. Domb and Pagel (2001) subsequently found that degree of ornamentation in olive baboons covaries positively with lifetime female reproductive success. Furthermore, males compete for and groom highly ornamented females more than they do their less ornamented counterparts.
We suspect that Pagel’s ideas, with some modification, are correct about many sexual swellings in females. We recognize, however, that these ideas have been controversial (e.g., Nunn, 1999; Nunn et al., 2001; Zinner et al., 2002; Zinner et al., 2004). We suggest that some of the questions about Pagel’s hypothesis are due to misunderstandings of its implications. Other criticisms can be deflected by modification of the hypothesis. At the same time, we fully acknowledge that much additional research testing these ideas, as well as a variety of alternative hypotheses for sexual swellings, is needed (Deschner et al., 2004; Dixson, 1998; Dixson & Anderson, 2004; Emery & Whitten, 2003; Hrdy, 2000; Nunn, 1999; Setchell et al., 2006; Setchell & Wickings, 2004; Stallmann & Froehlich, 2000; Zinner et al., 2004).
Just as female signaling of fertility status should be observed only under restrictive conditions, so should female ornamentation advertising quality—albeit conditions not uncommonly found. To be favored by selection, female ornamentation must accrue benefits to offset its costs. Males deliver those benefits, and they do so in currencies of nongenetic material benefits and genetic benefits (better genes) for offspring. Through signaling, females may be able to obtain greater benefits from males than they would otherwise. They can do so, however, only if males are constrained to have to make choices about which females to deliver benefits to—that is, when they cannot (or it does not pay for them to) provide equal benefits to all females. Under those conditions, high-quality females may be able to entice males to provide benefits to them and their offspring rather than to female competitors by signaling their quality. The greater the extent to which males must decide to whom to direct benefits, the greater the potential for female ornaments to evolve. And the greater the extent to which females differ in their phenotypic and genetic quality (e.g., as a result of disease prevalence or differential accrual of resources), the greater the potential for female ornaments to evolve. Put otherwise, selection can favor female advertising of quality only when males pay substantial opportunity costs if they direct mating effort or parental effort toward low-quality females and their offspring that they could more profitably direct toward trying to mate with or parent the offspring of high-quality females (either currently or in the future). These conditions, of course, are also the conditions under which male mate choice evolves (e.g., Bercovitch, 2001; Bonduriansky, 2001; Pagel, 1994; Parker, 1983; Shine et al., 2003).
The circumstances in which these conditions are most clearly met are ones in which males deliver important fitness-affecting, nongenetic material benefits to females, in which these benefits are not shared across females (e.g., females cannot collaterally benefit from male efforts to assist other resident females), and in which females exhibit substantial variation in phenotypic and genetic quality. In those circumstances, males may be limited in the amount of valuable resources or services they can accrue or deliver to females. And, even if they can deliver services (e.g., physical protection for offspring) to all, it may pay males to differentially allocate those services across females as a function of females’ quality (or quality of their offspring). Sex-role-reversed species, in which males provide greater parental investment than do females, are obvious examples. Still, sex-role-reversed species are uncommon (e.g., Andersson, 1994; Gwynne, 2001; Thornhill & Alcock, 1983). In most species in which males provide material benefits to females, female ornamentation nonetheless remains absent, perhaps because thresholds on the conditions that favor female ornamentation are not often reached.
Other circumstances in which female ornamentation may evolve are (1) when dense overlap in female fertile periods causes males to have to choose between multiple fertile females to pursue or (2) when even dominant males risk injury in male-male competition for females, particularly when females vary substantially in quality. In such circumstances, high-quality males may be selected to forgo mating with lower quality females because, under (1), they are forced to make choices through time constraints or because, under (2), the costs of competition (and loss of future reproduction) are not offset by the benefits of additional matings with lower quality females currently. In either set of conditions, female signals of quality could be correlated with the fitness of offspring and thereby evolve. (Under (2), females who signal quality mate with dominant males—but dominant males would presumably also win competitions and mate with them even if females displayed no quality-discriminating signal. Their advantage stems from the fact that females who cannot signal quality mate with less dominant males than if there were no signal. Hence, the offspring of quality signalers do have better genes than the offspring of female competitors, which leads genes for the signal to spread.)
Pagel (1994) proposed a form of the latter hypothesis to explain female sexual swellings. Specifically, he proposed that, because male mating effort in many Old World primates is costly (largely due to costs associated with aggressive combat), males can evolve to choose how to allocate their mating effort based on a signal of female quality. Females with preferred signals hence obtain genetic advantages for offspring. In fact, though this scenario appears to be plausible, whether female signals of quality actually have evolved in absence of male delivery of nongenetic material benefits conditional on female signal size is an unanswered question. (Though the male baboons in Domb and Pagel’s [2001] study apparently provide little to no paternal care, females may nonetheless derive material benefits from mating with dominant males, such as reduced levels of aggression toward offspring.) As we show, the nonmutually exclusive hypotheses that, on the one hand, female signals of quality derive benefits through male delivery of material benefits or, on the other hand, female signals of quality derive benefits through genetic benefits for offspring by regulating male mating effort offer some different predictions.
Female ornamentation can evolve in a variety of different mating systems, as illustrated in Old World primates. Female sexual swellings are most commonly observed in species living in multimale, multifemale groups (Clutton-Brock & Harvey, 1976; Dixson, 1998; Hrdy, 1997; Nunn, 1999). Indeed, the phylogenetic record shows that they tend to evolve in species that have these social structures and have multiple origins in the phylogenetic tree (e.g., chimpanzee swellings originated in their lineage after it diverged from ancestors shared with humans and gorillas and independently of its evolution in baboons and macaques; e.g., Dixson, 1998; Pagel & Meade, 2006). In some of these species, males care for their own genetic offspring more than for unrelated juveniles. In addition, males protect estrous mates from other males, including those that may sexually coerce estrous females (e.g., Maestripieri & Roney, 2005). Males delivering material benefits to females should be selected to prefer high-quality over low-quality mates, due to cost to the male of providing the benefits (e.g., a male’s lost time foraging associated with his mate guarding, as in baboons; Alberts, Altmann, & Wilson, 1996).
Yet female sexual swellings are found in other primate mating systems, as well (although these swellings are generally of smaller size and thus, presumably, are less costly). In a langur monkey, which lives in spatially separate, polygynous one-male groups, swellings may have evolved through intense competition for disproportionately large shares of the harem master’s time-limited material services and delivery of benefits (Tenaza, 1989). Females of a socially monogamous gibbon also display them (Dahl & Nadler, 1992). These females may compete for pair-bond partners who are good providers or material benefits from extra-pair males. Females of New World primate species in which males invest considerably in offspring (e.g., marmosets and tamarin monkeys) lack sexual swellings, but may have evolved alternative ornaments for signaling quality, notably through scent displays (see Heymann, 2003).
In some species of Old World primates, female swellings are particularly pronounced and, as a result, costly (see brief review in Emery & Whitten, 2003). Nunn (1999) found evidence that exaggerated female sexual swellings in primates are associated with three conditions. First, primate species with large sexual swellings live in groups with more males than do those without large sexual swellings. We suggest that this pattern arises because females in groups with many males face relatively great risks of aggression toward offspring and, hence, gain considerable material benefits by mating with males who can defend (and will not aggress against) their offspring. Second, female primates with large sexual swellings have sex for longer periods of time. In our terms, they have longer periods of extended female sexuality. As we argued in chapter 3, female extended sexuality, too, is favored when females can garner male material benefits through sex. In many Old World primates with sexual swellings, benefits of extended sexuality are achieved through reduction of aggression toward offspring via paternity confusion. Covariation between costly female signaling of quality and duration of extended sexuality is hence to be expected. We emphasize, however, that these adaptations are independent and require different explanations (e.g., costly signals of quality could evolve in absence of greatly extended sexuality). Third, exaggerated female sexual swellings tend to be found in species that lack seasonal breeding. By chance alone, females of seasonal breeders will tend to be fertile at the same time that other females are. Nunn (1999) argued that Pagel’s (1994) hypothesis should expect that seasonal breeding should favor costly signaling of quality to attract male attention and mobilize male mating effort; in seasonal breeders, females often come into estrus simultaneously and hence compete for dominant males’ attention when fertile. Thus, he argued, the relative absence of large swellings in these species is inconsistent with the idea that swellings are honest signals of quality. As we noted earlier, however, one must separate the honest signaling hypothesis that argues that benefits are achieved at least partly through delivery of male material benefits over a temporally extended period (including periods of time in which females do not display swellings) from Pagel’s specific hypothesis, which proposes that benefits are achieved solely by mobilizing male sexual interest when females are in estrus.
In fact, we suggest that honest signaling theory can account for the exaggerated ornamentation in nonseasonal breeding primates. Seasonal versus nonseasonal breeding is distinguishable partly on the basis of what most affects optimal timing of female reproduction: immediate availability of resources (for females or offspring), as it varies seasonally, or female condition and resource accrual (which may peak during any season). In species that breed seasonally, variation in female condition and resource accumulation matter less than variation in resources across seasons (a reason they breed seasonally). In species that breed nonseasonally, female condition and resource accumulation (including accumulation of social resources) matter more. In long-lived primates, females acquire, over the long term, nutrients, energy stores, social status, protective alliances with males, and other social resources, all of which may affect outcomes of female reproduction. Again, some of these factors probably matter more in nonseasonal breeders than in seasonal breeders, on average. And some of these resources (e.g., alliances) are delivered by males. Females of nonseasonal breeding primates, then, may be under stronger selection to signal their quality and condition to males, so as to accumulate male-delivered benefits.
Consider, for instance, female baboons (Papio). Females reproduce throughout the year, though successful reproduction may depend on ecological factors such as rainfall, ambient temperature, and group size (Beehner, Onderdonk, Alberts, & Altmann, 2006). Female baboons have moderate to large swellings. Swelling size may reveal to males when an individual female is in a condition favorable to successful reproduction (independent of season). Furthermore, through honest signaling of condition, females can gain some of the male-delivered benefits that further promote successful reproduction (e.g., through alliances with males). During their periods of sexual swelling, female baboons form alliances with males (Dixson, 1998). And, indeed, females often exhibit swellings well in advance of becoming consistently ovulatory in their cycles, purportedly to impress males and gain social alliances. (Female baboons may swell up to 18 times prior to conceiving; Altmann, Altmann, Hausfater, & McCuskey, 1977; see also chapter 6, this volume.) More generally, when nonseasonal reproduction reflects the fact that females of the species rely on accumulated social and other resources for successful reproduction, females may exhibit large sexual swellings, which function to honestly signal condition and thereby, when exaggerated, lead to acquisition of resources.
At the same time, nonseasonal reproduction is clearly not necessary for large female sexual swellings to be directly selected. Female Barbary macaques, for instance, are seasonal breeders and yet possess large sexual swellings (Mohle et al., 2005).
As we have acknowledged, female sexual swellings in Old World primates are often largest when their probability of conception is near peak within the estrous cycle (for review, see Dixson, 1998; also Zinner et al., 2004; Deschner et al., 2004). We have emphasized, too, that females also often exhibit swellings when infertile (e.g., Anderson & Bielert, 1994; Mohle et al., 2005). Again, we have argued that the link between swelling and estrus does not imply that females have benefited through signaling fertility per se. The link must nonetheless be explained. According to our view, it arises because at this time females optimally advertise quality and males optimally attend to signals of quality.
Consider signaling of quality from the female’s perspective. Females may benefit most from signaling when fertile for two reasons. First, at this time they may most benefit from male delivery of material benefits (e.g., protection from sexually coercive males). Second, at this time males attend to them and are most interested in mating with them, based on preexisting by-product cues of cycle-related fertility. Hence female efforts to signal quality are most likely to be recognized at this time. Relative to the traditional view, this idea reverses a key sequence of events in the evolution of signaling. In the traditional view, female ornaments lead to male attention to females at peak fertility. By contrast, the idea we propose is that male attention to females when near peak fertility leads females to display ornaments at that time (though, we emphasize, they obviously do signal at other times as well).
From the male’s perspective, too, female signaling of quality optimally occurs near peak fertility. At that time, males are most focused on fertile females. Competing demands on male attention are minimized when quality signaling occurs simultaneously with female fertility (for compatible arguments, see Pagel, 1994; see also Bercovitch, 2001; Domb & Pagel, 2001).
Finally, another potentially relevant fact is that female sexual swellings are estrogen dependent (Dixson, 1998). As we mentioned earlier in this chapter (and discuss in further detail in chapter 6), women’s estrogen level is highly sensitive to and related to energy storage and balance and, likely, to overall condition (particularly as it relates to readiness to reproduce). These associations may be true of Old World primates in general (Ellison, 2001). Hence the honest information provided by elaboration of female sexual swellings may be anchored fundamentally in the temporal correspondence between the fertile period of the cycle and females’ production of high estrogen levels. The fertile phase, then, may be the time when females can most validly display their quality to males. This is not to say that males depend heavily on swellings themselves to assess cycle fertility (or that female swellings were directly selected because they provide such information). Males primarily use by-products to assess cycle-related fertility status, just as males of mammalian species in which females lack swelling do. Rather, the argument is that males assess female condition and readiness to reproduce and that females can often best provide information about those qualities (through swelling displays) when they near maximum fertility.
Despite the common positive covariation between size of sexual swellings and cycle-relevant fertility status among female primates, individual females exhibit even stronger associations. Sexual swellings of individual female common chimps change across approximately 10 days of swelling (in a 36-day reproductive cycle), for instance, and are maximally enlarged during the 4 days of the cycle at which conception probability peaks. Dominant males respond by increased mating rate and mate guarding of highly fertile females (Deschner et al., 2004). One explanation of this pattern is the graded-signal hypothesis (Nunn, 1999). Females in species with swellings, Nunn suggested, face the dual problems of biasing paternity toward dominant males on the one hand and confusing paternity on the other. They solve these problems, he argued, by signaling maximal fertility when at peak fertility, thereby drawing attention and mate guarding from the most dominant males, and mating with submissive males when less fertile but still sexually swollen (albeit to a less exaggerated degree).
According to Nunn’s hypothesis, dominant males should use swelling size itself to decide whether to mate with females. In our way of thinking, fertility status per se, as reflected through by-product cues of fertility status (e.g., scents) and not swelling size, should largely affect male decisions about how to direct sexual effort. In fact, Deschner et al. (2004) found evidence for the latter, but not the former, in chimpanzees. Though, on average, females tend to show maximum swelling around ovulation, for individual females swelling may peak several days before or after peak fertility. Attention to by-product cues of fertility status, not swelling size, explains why and how dominant males focus sexual attention on and often monopolize females at peak conception risk. The same applies to long-tailed macaques, another species in which maximum sexual swelling does not reliably correspond to peak fertility in the cycle (though estrogen does): Males focus their mating effort on females in peak fertility per se (Engelhardt et al., 2004; Englehardt et al., 2005)
These findings regarding chimps and macaques are also inconsistent with one version of the swelling-as-quality-signal hypothesis—namely, Pagel’s hypothesis that swellings function to mobilize male mating interest. The findings are fully consistent, however, with the hypothesis that these signals of quality at least partly benefit females through male delivery of material benefits. Males’ attention to female fertility in the estrous cycle reflects male mating effort to inseminate females. It need not reflect any effort to offer particular females material benefits. In many circumstances, males should be sexually interested in fertile females who are not particularly of high quality. When multiple swollen females are at peak fertility, males may have to decide which females to pursue on the basis of swelling size per se. As well, even dominant males may also adaptively decide not to compete for copulations with low-quality females if competition risks injury.
Put otherwise, according to the hypothesis that female signals of quality function to obtain male delivery of material benefits, female swellings do not necessarily motivate male interest in copulation per se (cf. Deschner et al., 2004). They motivate male material benefit delivery to females. In chapter 6, we discuss exaggeration of sexual swellings in Old World primate adolescents, whereby subfecund females exhibit sexual swellings to a degree greater than older fecund females. These swellings do not stimulate male sexual motivation. Arguably, they instead function to attract delivery of material benefits from males.
This is not to say that males should not pay attention to female swelling size when deciding how to allocate mating effort (due to its costs), as well as delivery of material benefits. As just noted, they should. Sexual interest, however, need not be evoked by swellings per se.
As noted previously, Domb and Pagel (2001) found strong support for the hypothesis that swelling size signals female reproductive quality in a wild population of olive baboons. Swelling size positively predicted earlier maturation, lifetime number of offspring produced, and number of surviving offspring. (See Zinner et al., 2002, for discussion of study limitations.)
Emery and Whitten (2003) reported a positive association between female ovarian function (notably, estrogen levels) and swelling size measured in the same reproductive cycles of chimpanzee females. In humans, estrogen levels vary on a graded continuum with fertility of cycles; cycles in which estrogen is highest tend to be most fertile (e.g., Ellison, 2001; Lipson & Ellison, 1996; see also chapter 6, this volume). The same may be true of other primates. If so, swelling size may be an indicator of current reproductive condition. Although most of the comparisons that Emery and Whitten (2003) had available were between individual females, they also found evidence suggesting that, across different cycles of the same female, swelling size predicts ovarian function. Consistent with this claim, Deschner et al. (2004) found that, during cycles close in temporal proximity to an actual conceptive cycle of the same female, female chimpanzees had larger swellings than they developed during cycles several months prior to an actual conceptive cycle (cf. Setchell & Wickings, 2004, who did not find a difference in swelling size across conceptive and nonconceptive cycles in mandrills). Naturally, a valid indicator of quality or condition may not only discriminate between females but also may discriminate reproductive condition as it varies across cycles within individual females. Hence, these findings are consistent with a version of the quality-signal hypothesis.
That need not mean, however, that swelling size cannot also reveal long-term individual differences in ability to reproduce (Emery & Whitten, 2003). Based on observations of a small number of females of a semi-free-ranging group of mandrills, Dixson and Anderson (2004) argued that, due to physiological challenges, low-ranking females experienced difficulty maintaining estrogen levels sufficient to trigger ovulation, as revealed by swellings. A larger (but still small) study of the same group yielded mixed findings (Setchell & Wickings, 2004). Though results hinted that males compete more for females with large swellings, swelling size did not reliably relate to quality variables, such as the mean number of cycles to conception. A subsequent study that examined disease loads and health indicators also yielded ambiguous results (Setchell et al., 2006). Of 20 correlations between swelling size and parasite loads, 18 were in the predicted negative direction. As sample size was only 10, however, no individual correlation was statistically reliable.
Critics have claimed that Pagel’s hypothesis is challenged by a number of additional findings: Males are sexually attracted to fertile females, not females with the largest swellings; infertile females (e.g., pregnant females) sometimes display swelling, and in some species adolescent females have the maximal swelling size; exaggeration of female swellings does not coevolve with increased male canine size (see Nunn et al., 2001; Zinner et al., 2002.) Our proposal—which, unlike Pagel’s, does not argue that swellings function to incite competition between males—is not challenged by these findings. Moreover, because swelling size may vary across females’ life history in patterns different from their age-specific fecundities (e.g., young females may have maximal swelling size prior to having reached peak fertility), a lack of association between this purported quality indicator and female reproductive success in short-term studies (e.g., Setchell & Wickings, 2004) is not inconsistent with our proposal. Neither is a lack of association between swelling size and immediate sexual interest (see chapter 6 on the function of adolescent swellings). Domb and Pagel’s (2001) work remains the only study that has examined associations between female swelling size and lifetime reproductive success in any species, and it found strong associations.
Clearly, additional studies are needed.
In the world of animals, female sexual ornaments are widespread. Based on theoretical arguments, we suggest that they are typically honest signals of female personal quality and condition, not of fertility within females’ reproductive cycles—and this interpretation applies specifically to female sexual skins and swellings of Old World primates. Males are expected to be under strong sexual selection to find fertile females, and, hence, costly signals of fertility within the cycle should be rare. Generally, cues that females emit when fertile in their cycles have been interpreted as signals designed to assure insemination. One version of this notion claims that females signal to incite male–male competition, which assures insemination by males with good genes. Females typically do not have signaling adaptations that function to ensure insemination by males in general, though they do have other adaptations that function to improve female choice. Female primate copulation calls may well be adaptations, but are not fertility signals or adaptations to incite competition per se.
Conditions that favor the evolution of female ornaments include circumstances in which females vary in quality and females can obtain nongenetic material benefits from males who discriminatively deliver those benefits. Variation in these conditions may explain the distribution of sexual swellings and other ornaments (such as scent markings) among nonhuman female primates. Female sexual swellings may tend to be large when females are at peak fertility in their cycles because females may optimally signal quality and males may optimally assess females’ quality at this time. When the size of a female’s sexual swelling does not exactly peak when fertility in her cycle is maximal but the presence of by-products of female fertility (such as estrogen-related scents) does, male sexual interest is best predicted by actual fertility status, not swelling size. Furthermore, in some nonhuman primates, adolescent females display maximal swellings but do not attract male sexual interest (see also chapter 6). These patterns are consistent with the view that swelling size is not a primary determinant of male mating interest but rather functions as a signal of individual quality to gain material benefits from males.