The Princeton Guide to Evolution

Speciation and Sexual Selection

Janette W. Boughman

OUTLINE

1. Can sexual selection generate diversity?

2. Patterns of speciation by sexual selection

3. The mechanisms of sexual selection that cause speciation

4. Sexual selection and postmating isolation

How does the diversity of life arise? Many of the most striking differences among species occur in traits involved in mating—especially traits that males use to compete with other males or to attract females, or that females use to select mates. Big differences in these same traits also make mating between species unlikely, contributing to reproductive isolation. It seems fairly intuitive, then, that whatever process causes differences in mating traits is involved in the formation of new species. The most likely process is sexual selection, which is defined as variation in mating success among individuals varying in phenotype within a population, caused either by males competing for access to females or females preferentially mating with some males over others. Sexual selection might be able to explain why some groups of organisms have many species—hundreds of thousands of beetles in a single order of insects—while others have very few—only five species of horseshoe crabs represent an entire class. But is sexual selection the only, or even the most important, process creating new species? How do we find out? One approach is to use information in the patterns of diversity of mating traits and species richness. If sexual selection is important in speciation, then groups of organisms that experience sexual selection should have more species than those that do not. Another approach is to explore how sexual selection causes mating traits to diversify and to ask whether differences in mating traits actually keep species from mating with each other. Results from these studies can be compared with those investigating natural selection in speciation to determine their relative roles. These are the lines of evidence considered in this chapter.

GLOSSARY

Allopatry. Geographic distribution in which populations or species are completely separated, with no contact and no opportunity for gene flow.

Antagonistic Coevolution. Coevolution in which the evolutionary interactions between two parties (two sexes or two species) impose costs on each other because of different evolutionary interests. The metaphor of an arms race is often used to describe the process and its outcomes.

Assortative Mating. Mating between individuals that are similar in a trait or set of traits, such as size assortative mating in which large males mate with large females and small males with small females. Also used to indicate preferential mating between individuals of the same species over individuals of other species.

Biological Species Concept. The classification of species as groups of potentially interbreeding natural populations reproductively isolated from other such groups.

Coevolution. The process in which evolutionary change of one species influences the evolution of another species.

Dimorphism. Differences between males and females of a species in size, structure, color, ornament, or other morphological trait(s), not including the sex organs.

Fisherian Runaway Sexual Selection. A model of sexual selection conceived by R. A. Fisher to explain the exaggeration of both male display traits and female preferences in the absence of benefits to females. Genetic correlation between male trait and female preference creates continual evolutionary exaggeration of both until the reproductive benefit of the exaggerated male trait is balanced by the cost of producing it.

Gene Flow. Movement of genes from one population or species to another through mating between individuals of those populations or species.

Good Genes Sexual Selection. A form of sexual selection in which females obtain genetic benefits, or “good genes,” from mating with particular males. Female preferences evolve for male traits that indicate genetic benefits.

Local Adaptation. Evolutionary change to increase fitness of organisms in the local environment. Locally adapted individuals or populations have higher fitness in the local environment than the source (ancestral) population or other populations experiencing different conditions.

Mate Preference. Selection of mates based on criterion values of specific trait(s). Preference influences the propensity of individuals to mate with certain phenotypes and consists of two components: preference functions describe the way in which trait values are ranked, and choosiness is the effort an individual invests in making its choice.

Mating Trait. Secondary sexual traits involved in mating. These include display traits and competitive traits for males, and mate search and preference for females.

Reproductive Isolation. Speciation that occurs via the evolution of isolating barriers, which are characteristics of organisms that keep individuals in one population from exchanging genes with other populations. Reproductive isolation can occur by preventing individuals of separate species from mating (premating isolation) or by selecting against hybrids (postmating isolation).

Sensory Drive. A mechanism of sexual selection in which mate preferences evolve as a by-product when communication systems adapt to local conditions. A communication system includes sensory adaptations, preferences, and signaling traits. Although most often studied in the context of mating, sensory drive can occur for communication in any behavioral context.

Sexual Conflict. Occurrence of conflicting evolutionary interests and optimal strategies for reproduction between the two sexes, including aspects of reproduction such as mating rate. Sexual conflict can give rise to antagonistic coevolution of traits in each sex, including both behavior and morphology, that mediate such conflicts.

Sexual Isolation. A form of premating isolation in which the choosy sex of one population or species (usually female) is less likely to accept members of the other population or species as mates.

Sexual Selection. Variation among individuals with different trait values in the number of mates acquired and in overall reproductive success, measured as the number of offspring produced. Intersexual selection involves choosiness by one sex for mates of the other sex based on trait values (often called female choice). Intrasexual selection involves competition within a single sex for access to the other sex, frequently through contests (often called male-male competition).

Signal. A trait modified by selection to convey information and to influence the behavior of individuals receiving it. Signals can be in various modalities, including visual, auditory, olfactory, and tactile, and can involve specialized structures or be purely behavioral. Male display traits are signals.

Speciation. The process by which one or more species evolves from another via genetic changes and the evolution of mechanisms that restrict gene flow.

Sympatry. Geographic distribution of species in which at least part of the ranges of two species overlap, allowing individuals to encounter one another and making gene flow possible.

1. CAN SEXUAL SELECTION GENERATE DIVERSITY?

This chapter focuses on sexual selection as a force causing evolutionary change and speciation. Sexual selection is defined as a process that arises from differences between individuals in the ability to attract mates or in the number of successful matings caused by differences in underlying traits. Males with traits that females prefer mate more often and leave more offspring; these offspring also inherit the traits of their (preferred) father. Over time, populations come to be composed primarily of trait values preferred by females. Because of the high importance of mating success and reproduction to overall fitness, sexual selection is likely to be strong. For reasons explored in this chapter, males of one species are very unlikely to have traits that females of another species prefer. The combination of strong selection and opportunity for mismatch of traits and preferences is one important reason that sexual selection is thought to lead to reproductive isolation.

Reproduction is a key element of an individual’s lifetime, and choosing a mating partner can have huge effects on overall fitness. The number of offspring an individual produces is the main measure of fitness, so even if an individual survives, without mating its fitness will be zero. Successful reproduction often involves mating displays typically produced by males, and preferences for those displays typically used by females to guide their choice of mates (see chapter VII.6). Male displays are often exaggerated adornments including patches of color and extravagant structures, or elaborate courtship behavior including dances, songs, and gift giving. Female preferences for one color over another, one song feature, or one smell over another can be quite extreme, leading the female to reject most suitors in favor of a particular male who meets her selective criteria. In many taxa, females have invested heavily in producing eggs and will continue to invest in parental care, so commonly, they do the choosing to find the best mate to father their offspring. Choosy females generate sexual selection on males, leading to evolutionary change in mating traits.

These considerations provide a second reason that sexual selection is thought to be important to speciation. Sexual selection is of fundamental importance to mating, and it causes mating traits to evolve and diverge, thereby influencing premating isolation. This chapter focuses on premating isolation, because sexual selection is especially likely to influence its evolution. Sexual selection can affect some aspects of postmating isolation as well, which is considered briefly at the end. A key form of premating isolation is termed sexual isolation, which occurs when mating preferences and/or display traits differ between populations, so that females of one population or species do not find males in the other attractive and hence refuse to mate with them. This assortative mating between species, or “like mates with like,” can reduce gene flow and enhance reproductive isolation. Sexual isolation is commonly the primary isolating barrier between species in nature. Because sexual selection influences the evolution of mating traits leading to sexual isolation, sexual selection likely drives speciation in nature.

A third reason that sexual selection is thought to be important to speciation is that frequently, the most conspicuous differences between closely related species are traits involved in mating. Examples of this pattern abound. Male birds of paradise have spectacularly elaborated plumage including elongated feathers and brilliant colors, and males use these plumes in odd courtship dances to attract highly choosy females. No two species are alike, and both the elaboration of plumage and courtship as well as the marked differences among species are due to strong sexual selection. When species differ substantially in mating traits but little in ecological traits, sexual selection is implicated as especially important. This characteristic is also seen in lacewing insects: although many species are morphologically and ecologically very similar they produce vibratory mating duets that differ substantially. Males and females who cannot duet properly because they “sing” a different vibratory song do not mate, and this is the key mechanism that isolates the species. Such patterns indicate that sexual selection plays a role in speciation.

Scientists used to think that speciation would take a very long time to occur and that they could understand it only by looking back in time for millions of years (see chapter VI.1). But it turns out that we can witness speciation in action. In some cases, new species arise over tens or hundreds of organismal generations, and within a human lifetime: speciation is happening all around us. Thus we can look through the window this opens on the process of speciation and study what makes it tick. Rapid speciation (speciation in action) is especially likely when sexual selection or natural selection is involved, because these evolutionary processes typically increase the rate of evolution. In one example, Susan Masta and Wayne Maddison studied a group of jumping spiders found in the “sky islands” of Arizona (mountain ranges isolated by intervening desert) that diversified as recently as 10,000 years ago. Males sport intricate color patterns—swaths, stripes, and splotches of red, yellow, blue, black, and white on their face and forelegs—that differ markedly among populations inhabiting different mountains. Males use their facial and foreleg “paint” to attract females, indicating that sexual selection plays a role in the extreme differences among populations in color pattern, which is leading to rapid speciation. Speciation is not yet complete and has proceeded further in some populations than others. Therefore, pairs of populations that are strongly reproductively isolated can be compared with those less isolated to test specific mechanisms causing the evolution of reproductive isolation.

In recent years, research has gone past the initial step of describing differences in mating traits between reproductively isolated species to answering other pressing questions about when and how sexual selection causes reproductive isolation. We now have insight into the forms of sexual selection involved and the way in which sexual selection and natural selection interact to cause divergence. There are two complementary ways to address the role of sexual selection in speciation, covered respectively in the next two sections. First, we can look for patterns of diversity consistent with the action of sexual selection. Second, we can study the process of sexual selection and ask if it generates diversity.

2. PATTERNS OF SPECIATION BY SEXUAL SELECTION

Does sexual selection lead to higher diversity? This question can be addressed by testing for correlations between the presence or degree of sexual dimorphism (or other proxies for sexual selection) and measures of taxonomic diversity, such as the number of extant species. By testing such a correlational relationship in many taxonomic groups, one can assess whether sexual selection is generally associated with higher diversity, as would be expected if it promotes speciation across the tree of life. For example, Nathalie Seddon and colleagues showed that phylogenetic groups of antbirds that are highly dichromatic (male and female plumage colors differ) are more species rich than antbird groups that have little or no dichromatism. Many other comparative studies have found positive associations between measures of sexual dimorphism and species richness, supporting the idea that sexual selection causes speciation. However, not all studies have found this pattern, raising some doubts about the role of sexual selection in driving speciation generally. One reason for the discrepancies among studies may be that proxies for sexual selection like dimorphism are inexact ways to measure sexual selection, or that sexual selection is an important driver of speciation in some groups but not in others. Another possibility is that these studies typically do not take into account species extinctions. Counting the number of existing species in a phylogenetic group reflects not just speciation rate but also extinction rate. If sexual selection influences both, then its role in initiating speciation would be obscured. By addressing these possibilities and evaluating the relationship across many comparative studies, Kraaijeveld and colleagues calculated an overall estimate of how strongly sexual selection correlates with species richness in many groups of animals. The correlation is positive but low (correlation value of 0.07 to 0.14), which suggests that sexual selection does indeed contribute to speciation in many groups of animals, but that it does not act alone. Interestingly, this correlation is similar in magnitude to that found by Daniel Funk and colleagues between measures of ecological divergence and reproductive isolation, which hints that sexual and natural selection may contribute equally to speciation.

Does sexual selection increase the amount of reproductive isolation? To answer this question, one can compare the extent of reproductive isolation between closely related species with the strength of sexual selection between them, and with the degree they have diverged in sexually selected traits. The logic is that strong (and divergent) sexual selection causes a large amount of divergence in sexually selected traits between closely related species, which then results in reproductive isolation. For example, pheasants, whose species differ substantially in the traits males use to attract females, are expected to exhibit more reproductive isolation than parrots, whose species have much smaller differences in sexually selected traits. The relative importance of sexual selection to speciation is evaluated by comparing the amount of divergence in sexually selected traits and female preferences to that in traits thought to have diverged by natural selection instead. These divergence metrics are then correlated with reproductive isolation to suggest which force, natural or sexual selection, is more important. Key findings from such comparisons are that the amount of difference among species in male mating traits and female preferences predicts the extent of reproductive isolation better than the strength of female preferences for the desired male trait within species. These results suggest that sexual selection by itself is not enough to cause reproductive isolation, even if it is strong. Sexual selection needs to be divergent to cause new species to form.

3. THE MECHANISMS OF SEXUAL SELECTION THAT CAUSE SPECIATION

In theory, female preferences can generate sexual selection on traits by several mechanisms, with some more likely than others to cause divergence between populations in those traits, and ultimately to generate sexual isolation. The mechanisms of sexual selection include sensory drive, good genes, Fisherian runaway, and sexual conflict. Which of these are most likely to cause divergence in mating traits? The mechanisms can be grouped into two general classes according to underlying causes: differences between the environments occupied by populations (sensory drive and good genes), and interactions between the sexes (Fisher’s runaway and sexual conflict). The environmentally dependent processes (sensory drive and good genes) lead to predictable associations between mating traits and environmental differences. In contrast, interactions between the sexes that generate divergent sexual selection by Fisher’s runaway process or by sexual conflict have little to do with environment and generate evolutionary change in arbitrary, unpredictable directions. Whether environmentally dependent mechanisms are more or less likely to cause reproductive isolation than the arbitrary mechanisms is the subject of ongoing research.

Sensory Drive: Local Adaptation of Communication Systems as a Cause of Sexual Isolation

Sensory drive is a process by which some aspect of the sensory world predisposes individuals to attend to and prefer particular features of communication signals. Sexual selection through sensory drive is essentially a hypothesis about the effect of environment on shaping sensory systems, the preferences that depend on senses, and the display traits or signals that are preferred. Animals rely on their sensory systems to acquire information on predators, prey, and mates, and a sensory system that works well in one environment may not be so effective in another. This means that populations in different environments are likely to evolve differences in details of their sensory systems. These differences might include which senses they rely on most (e.g., vision for daytime visual predators, and hearing for nocturnal animals), and how those sensory systems are tuned, for example, what sound frequencies they evolve to hear best, what colors they evolve to see best, what smells they evolve sensitivity to. For example, deep in the ocean the prevailing light is blueshifted because water absorbs red wavelengths. Species of snapper fish that live in deep water have evolved to see blue light much better than closely related snappers that live in shallow estuaries—the eyes of deep-water species are tuned to blue light. This sensory tuning is favored because heightened sensitivity to blue light helps the fish see and discern objects in their environment, whether those objects are prey, predators, refuges, or members of their own species. Sensory tuning can, in turn, affect which mating traits are preferred even when conscious choice is not involved; those that match sensory tuning are likely to be easier to detect, and evolution favors mate choice based on easy-to-detect mating traits. Thus, sensory systems influence mate preferences, which should result in sexual selection on mating traits to match sensory systems.

How well a signal is transmitted through the environment should also influence how well it is detected, owing to the physics of sound, light, or chemical diffusion. Signals that travel easily through one environment may be degraded in another because of the physical interaction of the signal with the environment. Degraded signals are likely to be harder to detect, and degradation may obscure features of the males’ signal essential for attracting females. In one example, populations of torrent-eared frogs that live near noisy rivers and waterfalls have evolved to produce louder and higher-pitched calls that can still be heard over the noise. Selection acts on signals to favor those that are well adapted to local transmission conditions.

Local adaptation of communication systems means that populations in different environments are likely to evolve differences in mate preferences and mating traits, which can lead to sexual isolation as a by-product of local adaptation of communication systems. Females in one population prefer trait values that males in the other population do not have and thus do not mate with them. Reduced mating arising from this mismatch between preference and trait among populations slows gene flow and leads to sexual isolation, pushing populations toward becoming distinct species. Sensory drive is thought to be widespread, potentially making it a very important way that sexual selection causes speciation.

Sensory drive causes sexual isolation in lake-dwelling threespine stickleback fish. Populations of two species that live in clear water see red well and prefer it. Males of these species display large patches of bright red color during the breeding season. Red is highly visible in clear water that has full-spectrum light. However, many lakes in the northern latitudes have tannin-stained, redshifted water because organic molecules in the water absorb blue wavelengths. The stickleback species that live in redshifted water do not see red as well, females do not prefer it, and males have evolved to display black color instead of red. In these redshifted habitats, the black males are contrasted against the red background light, making their black mating garb highly visible and easy to detect. Local adaptation of color vision and coloration matters to speciation, because the more two stickleback species differ in color preference and male color, the less likely they are to mate. This characteristic generates sexual isolation between species in different light environments and is an important factor in their rapid speciation. Sticklebacks are an example of speciation in action, as the distinct species arose in lakes that formed after the glaciers receded in the last ice age less than 15,000 years ago.

Additional evidence that sensory drive plays a pivotal role in speciation can be found by identifying the genes involved in sensory adaptation that also underlie reproductive isolation. This approach revealed that speciation in African cichlid fish occurs at least partly by evolution in genes that control color vision, and implicates sensory drive in the evolution of divergent female preferences for male color. Males of different cichlid species display either red or blue bodies to attract females, who prefer color patterns displayed by males of their own species; thus color differences enhance sexual isolation. Color vision depends on visual pigments called opsins found in specialized cone cells in eyes. Having different opsins imparts differences in how well individuals see particular colors, which has been shown to influence color preference and the strength of sexual isolation. In an elegant series of experiments, Ole Seehausen, Karen Carleton, and their colleagues (2008) found that differences in opsin genes underlie differences in color vision and color preference in several species of African cichlids from Lake Victoria. Opsins have evolved in response to the light environment, which varies among locations in the lake from clear to murky to redshifted. The fish have evolved to see the dominant water color best. These scientists have also shown that water color is correlated with the body colors males display to attract females—they display bright red in deep water, and blue in shallow water. When color signals and preferences differ among cichlid species, they are sexually isolated. This series of studies identified the sources of natural selection and sexual selection, their contribution to sexual isolation, and the genes that underlie female preferences in different species. Although only one particular set of cichlid species has been studied intensively, similar patterns of divergence in coloration with water clarity and depth are also seen in other groups of African cichlids, implicating sensory drive and color evolution as an important cause of speciation in this group. This is no small feat, as the African cichlids are a textbook case of extraordinary diversification. Hundreds of species have arisen within a very brief period in several of the large African Rift Valley Lakes.

Good Genes: Environmentally Dependent Benefits of Mate Choice and Locally Adapted Males

Sexual selection for good genes comes about because compared with other males, some males in a population carry superior alleles (good genes) that confer high fitness, such as alleles that help resist disease. Females who mate with these males will obtain these “good genes” for their offspring. Offspring of preferred males should have higher fitness than offspring from unpreferred males who do not have “good genes.” Mating preferences that help females select these superior males will be favored by natural selection, and sexual selection will favor male traits that indicate genetic quality. Key requirements for good genes sexual selection to occur are variation in male genetic quality coupled with male traits that honestly indicate that quality.

Until recently, good genes sexual selection was not thought to contribute to reproductive isolation, but we now know two ways in which it can play a central role. Both have to do with differences in environment, and both will generate sexual isolation between populations from different environments. In the first way, benefits that females derive from choosing particular males as mates can depend on the environments the females and their offspring inhabit. These are known as context-dependent benefits. Because benefits from mate choice guide preference evolution, context-dependent benefits can lead to different preferences for populations in different habitats. Alison Welch found that female gray tree frogs gain benefits from mating with males who produce calls of long duration when they live in low-density habitats but may pay costs for mating with these same males in high-density habitats. In frogs, large size is advantageous, but so is rapid development. Offspring of long-calling males are larger at metamorphosis in low-density habitats (an advantage) but take longer to mature in high-density habitats (a disadvantage). This feature is expected to lead to different preferences by females in those two habitats. Females in low-density habitats should prefer long-calling males so that their offspring will grow large, but females in high-density habitats should avoid mating with long-calling males because their offspring will mature slowly and be at a disadvantage.

In the second way that good genes sexual selection can play a central role, locally adapted males have the particular alleles that make them well suited to local conditions; therefore, they have “locally good genes.” Being locally adapted means that males are well suited to feed on local foods, avoid locally abundant predators and parasites, and deal with the local climate; thus they are likely to be in good condition. In many cases, mating traits are bigger or brighter when males are in good condition; this characteristic is known as condition-dependent male trait expression. Females can gain benefits by mating with these locally adapted males because their offspring will inherit the locally adapted alleles and thus have high fitness. Moreover, females who prefer male traits that are condition dependent are more likely to choose these locally adapted, high-condition males, primarily because those locally adapted males will display the bigger and brighter traits that females prefer. In contrast, males who have immigrated from other populations will not have the locally beneficial alleles likely to be in poor condition, and as a consequence, to display small or dull traits. Local females will not prefer to mate with them, and female rejection of nonlocal males will create sexual isolation between the local population and more distant ones, especially when environmental conditions differ. Sander van Doorn, Pim Edelaar, and Franz Weissing developed these ideas in a theoretical model in 2009. The model awaits empirical tests, but this scenario may prove to be widespread.

Fisherian Runaway: The Role of Arbitrary Divergence of Mating Traits in Speciation

Fisher’s runaway process has the potential to rapidly amplify differences between populations in both maletraits and female preferences and, by doing so, to cause sexual isolation. Because coevolution between male trait and female preference is not dependent on environmental differences, it can occur in a multitude of directions. Allopatric populations are therefore likely to evolve in different directions even if they occur in similar environments. The direction of divergence is determined by chance factors such as the mutations present in the populations or differences in starting allele frequencies. With divergence possible in many arbitrary directions and the possibility for rapid change in mating traits, initially it seemed very likely that Fisher’s runaway could lead to sexual isolation between populations. Many theoretical models of sexual selection and speciation are built on Fisher’s runaway process, probably because the model is simple, as it does not assume natural selection on the female preference, and because the theoretical structure is elegant. Few empiricists, however, think that Fisher’s runaway by itself is likely to be responsible either for the evolution of female preference or for divergence in mating traits sufficient to cause sexual isolation in nature. A primary reason for this doubt is that Fisher’s process is not expected to lead to runaway evolution when females experience costs for being choosy. Costs such as increased search time, increased exposure to predators, or increased chance of remaining unmated seem quite likely to exist in nature, although these costs of choice are difficult to measure. When choice is costly and there are no compensating benefits, the expected evolutionary outcome is weak or no preference, and little or no exaggeration of male traits, neither of which is likely to enhance sexual isolation.

Some have suggested that Fisher’s runaway process might cause speciation in sympatry, where it is envisioned to cause a single population to split along two paths that evolve in a runaway fashion in different directions. This would occur if there are females in a population that strongly prefer quite different male traits, such as when some females prefer orange color and others prefer black color, or when some females favor complex song and others prefer simple song. Each type of female would select males they prefer, causing sexual selection and evolution in the male trait. Because of the genetic correlation established between trait and preference, the female preference would evolve in concert with the male trait, causing the male traits to evolve ever more exaggeration in two directions at once. Moreover, females who prefer black would be unlikely to mate with orange males, and vice versa, generating reproductive isolation between the diverging subpopulations. In this way the initial population containing both black and orange males with their black-preferring and orange-preferring females would end up as two separate reproductively isolated sympatric populations of all black or all orange. At this point the population will have split into two distinct species. This scenario requires disruptive sexual selection on male traits, which is created by the different female preferences. However, it also requires disruptive natural selection on the female preferences, which may be uncommon in nature. Female reproduction is limited primarily by the ability to acquire sufficient resources to reproduce, and many scientists think this is unlikely to generate disruptive selection on the preferences they have for male traits. Even so, new theoretical work is exploring these possibilities. The special combination of factors needed may be one reason why sympatric speciation via Fisherian sexual selection is probably rare.

The genetic correlation between male traits and female preferences central to the Fisher process is nevertheless likely to occur whenever there is sexual selection by female choice, even under scenarios involving good genes or sensory drive, because the genetic correlation arises inevitably from nonrandom mating, which occurs whenever females prefer some males over others. For example, under a good genes scenario, sexual selection is predicted to lead to a genetic correlation between female preference and male trait. Moreover, the male trait will be correlated with offspring fitness because preferred males also possess good genes; in this case their offspring will inherit a suite of genes: the fitness-related “good” genes, those for the elaborated male trait, and those for the stronger female preference. In practical terms, this means that detecting the mere presence of a genetic correlation between male trait and female preference in a natural population does not imply that the Fisher process is behind the evolution of exaggerated male traits, because every known process of sexual selection is predicted to lead to such a correlation.

Sexual Conflict: Coevolutionary Dynamics, the Mating Dance between the Sexes, and Divergence

Even though successful reproduction increases fitness for both males and females, the sexes achieve high fitness in different ways, which leads to different evolutionary interests of the sexes and can result in conflict over reproductive strategies and outcomes. The best outcome for males is likely to impose costs on females, and the reverse is also true. Termed sexual conflict, this feature can lead to rapid and dynamic evolution of male and female reproductive traits. The sexes coevolve, but do so antagonistically. The evolutionary response in one sex ameliorates the costs it experiences, but this adaptation imposes costs on the other sex. This can cause cyclic changes or escalations as adaptations and counteradaptations evolve.

For example, male water striders benefit by mating often and have therefore evolved behavioral strategies and morphological structures to increase mating rate, such as persistent courtship behavior and grasping structures. Female water striders need to mate only once, and when mating rates are high, females experience costs from male harassment, exposure to predators during courtship and copulation, and reduced time spent feeding. The result is selection for traits in females to resist frequent mating, such as high choosiness, evasive behavior, and morphological structures that resist grasping. When females are at their optimum mating rate (the mating rate that yields the most offspring over their lifetime), males are not, and vice versa. This difference generates continual evolutionary change. The antagonistic nature of coevolution also results in negative genetic correlations between male and female mating traits, and between male and female fitness, in contrast with the positive correlations expected with good genes sexual selection.

These interactions between the sexes generate strong sexual and natural selection: they cause male traits to evolve quickly in response to female resistance to persistent males, and females to evolve quickly to counteract these male adaptations. Sexual conflict thus creates dynamic evolutionary change in each sex in response to the other. Males and females are engaged in an evolutionary dance, often termed an arms race. Crucially, allopatric populations of the same species are likely to follow different evolutionary trajectories as a consequence of differences in ecology or genetic variation and the strong, dynamic selection imposed by each sex on the other. For example, if a mutation occurs in one population that alters the structures water strider males use to grasp females, but in another population a mutation occurs that increases male aggression, the counteradaptations of females are likely to be different in the two situations (perhaps exaggerated morphology in the first case and exaggerated behavior in the second). The result is that the two populations diverge; they will evolve different adaptations in males to increase mating rate, and different adaptations in females to reduce it. Similar to Fisher’s runaway, this coevolution can proceed in a multitude of directions, determined by the particulars of the populations involved. As a by-product of the coevolutionary arms race, sexual isolation can evolve because males and females from different populations are mismatched. However, the alternative outcome is also possible if males in one population are better able to overcome resistance from females in another population because those females have not evolved resistance to their persistence mechanisms. Therefore, whether sexual conflict promotes or inhibits sexual isolation is a matter for debate. This uncertainty is echoed in theoretical and empirical work on the subject.

Theory initially suggested sexual conflict would be likely to cause rapid speciation; however, those early claims have been modulated by later theoretical findings, which found that the conditions required for speciation were restrictive, and sexual isolation was rarely the predicted outcome. Several comparative studies of groups of species have indicated that sexual conflict is indeed associated with increased diversification rates or higher species richness, at least in insects. However, evidence is mixed from studies that experimentally manipulated the presence of sexual conflict in laboratory populations and followed evolutionary change over successive generations. Some experiments found that sexual isolation accumulated over time, but other experiments found no increase in sexual isolation. The jury is still out on whether sexual conflict causes speciation, but this active research area hopes to provide answers soon.

4. SEXUAL SELECTION AND POSTMATING ISOLATION

Sexual Selection against Hybrids

Even though sexual selection is primarily involved in the evolution of premating isolation, it may also be involved in postmating isolation, defined as reductions in hybrid fitness after hybrids are formed. For example, postmating isolation occurs if females discriminate against hybrid males. If hybrids are relatively rare, hybrid males will compete with males of the parent species for mating opportunities. The hybrids are likely to have intermediate, mixed or incompletely expressed phenotypes. If females of both parental species consequently find those hybrids unattractive, the hybrid males will have low reproductive success. Thus sexual selection against hybrids will limit gene flow even after hybrids have formed. Hybrid male chorus frogs are an example. The calls of males from the two parental species differ in pulse rate: one species gives calls with a more rapid pulse rate than the other, and females of each species prefer their own males’ pulse rate. Hybrid males call at an intermediate pulse rate and are not attractive to females of either parental species. The strength of sexual selection against these hybrids is substantially stronger than natural selection, indicating that sexual selection can, at times, be an important part of postmating reproductive isolation.

Sexual Selection and Genetic Incompatibility

Another possible way in which sexual selection can contribute to postmating isolation arises from genetic changes it causes between populations. Sexual selection causes rapid evolutionary divergence, and the genetic changes that occur in different populations may be incompatible when combined in hybrids. The alleles present in one population work well on that genetic background. If they did not, they would be selected against and not rise in frequency in the first place. However, those alleles have not been tested on the genetic background of other populations and may be incompatible with the genetic background of foreign populations. If this incompatibility occurs, hybrids may not survive or may have diminished reproductive success. This postmating isolation will restrict gene flow, and is termed Dobzhansky-Muller incompatibility. The role of sexual selection in this process is unknown at present, and there are no good examples yet from nature; however, given the rapid evolutionary change that sexual selection is expected to generate, it may be a fruitful avenue of future research and deserves more study. This and other genetic mechanisms of reproductive isolation are considered thoroughly in chapter VI.8.