The Princeton Guide to Evolution

VI.4

Speciation and Natural Selection

David B. Lowry and Robin Hopkins

OUTLINE

1. Types of natural selection contributing to reproductive isolation

2. Types of reproductive barriers and the effect of selection on their evolution

3. Considerations when studying natural selection and speciation

4. Reinforcement

5. Future directions

Natural selection is the process whereby heritable genetic variation changes in frequency as a result of its effect on survival and reproduction. The idea that natural selection plays an important role in speciation dates to Charles Darwin. Even so, major advancements in our understanding of how both ecological and reinforcing selection act to drive speciation have occurred since the mid-1990s. Extensive research investigating the role of selection in the process of speciation has revealed the importance of disruptive and directional selection in causing reproductive isolation between diverging groups of organisms. While the idea that ecological adaptation can cause reproductive isolation is basic, the way in which this process occurs is complex and often involves many agents of selection and multiple reproductive isolating barriers.

GLOSSARY

Assortative Mating. The preferential mating among individuals within a group of organisms based on similarity of phenotype.

Directional Selection. Natural selection favoring one end of the phenotypic spectrum over the other end of the spectrum.

Disruptive Selection. Natural selection favoring extreme phenotypes; intermediate phenotypes are the least favored.

Ecotype. A population or group of populations that have evolved a consistent suite of adaptations in response to local environmental conditions.

Natural Selection. The process whereby heritable genetic variation changes in frequency as a result of its effect on the fitness of an organism.

Parallel Speciation. The independent evolution of the same type of reproductive isolating barriers in response to similar agents of selection.

Reinforcement. The process whereby reproductive isolation increases as a response to natural selection against maladapted hybrids.

Reproductive Isolating Barrier. An evolved difference that acts to reduce the exchange of genetic material between populations, ecotypes, or species.

The main thesis of Charles Darwin’s On the Origin of Species is that natural selection, as opposed to special creation, is the cause of species diversity on planet earth. Although Darwin made this argument, the details of how he envisioned natural selection contributing to speciation are unclear and often contradictory throughout his writings.

Soon after the publication of The Origin, other evolutionary biologists, most notably Alfred Russel Wallace, explicitly argued that ecological adaptation plays a role in the formation of new species. However, by the late nineteenth century, research focused on linking adaptation and speciation had waned. Interest was revived during the 1920s when Göte Turesson published a flurry of papers, in which he coined the term ecotype. Often overlooked, Turesson’s research and theories inspired biologists of the time, particularly botanists, to conduct research investigating the connection between ecological adaptation and species formation.

Theodosius Dobzhansky’s Genetics and the Origin of Species, published in 1937, and Ernst Mayr’s articulation of the biological species concept (BSC) in 1942 defined speciation by the evolution of reproductive isolating barriers. Examples of reproductive isolating barriers include differences in mating preference between species, hybrid inviability, and hybrid sterility (see chapter VI.1). The BSC was a crucial development, because it directed researchers to focus on the mechanisms underlying the formation of reproductive isolating barriers to understand the process of speciation. With his third edition of Genetics and the Origin of Species in 1951, Dobzhansky articulated a list of ways in which natural selection could result in the formation of reproductive isolating barriers. In the same year, the botanist Jens Clausen published the book Stages in the Evolution of Plant Species, which assembled extensive evidence supporting the role of natural selection in the origin of many plant species. While Dobzhansky and Clausen described mechanisms by which natural selection could be important for speciation, they also argued that speciation often occurs through the accumulation of multiple reproductive isolating barriers arising from both selection and genetic drift.

After the 1950s, interest in the role of ecological natural selection in speciation diminished substantially. Studies of speciation became focused instead on the geography of speciation, especially the debate over the relative importance of allopatric, parapatric, and sympatric speciation (see chapter VI.3). Throughout the 1980s, it was thought that the major mechanisms involved in speciation were nonecological, often involving neutral genetic drift and chromosomal rearrangements.

Since the mid-1990s, the role of ecological adaptation in the formation of species has again become a major focus of evolutionary biologists. This research has been greatly enhanced by the feasibility of molecular techniques in diverse taxa. With modern tools in hand, researchers can now test important hypotheses about the role of natural selection in speciation.

1. TYPES OF NATURAL SELECTION CONTRIBUTING TO REPRODUCTIVE ISOLATION

Speciation usually results from the gradual accumulation of many reproductive isolating barriers. Natural selection can play a role in speciation when it leads to divergence between populations, and that divergence results in isolating barriers. Under most circumstances, selection does not directly favor an increase in reproductive isolation but rather reproductive isolation evolves indirectly as a by-product of selection. Reinforcement is the process whereby selection favors an increase in reproductive isolation to decrease the production of unfit hybrids.

The field of population genetics has defined three major categories of natural selection: disruptive, directional, and stabilizing selection. Both disruptive and directional natural selection are thought to play a major role in the formation of species and are discussed here in detail. Stabilizing selection occurs when intermediate phenotypes within a population are favored. The result of stabilizing selection is an increased frequency of individuals with the intermediate phenotype and a decreased trait variance in a population. Since speciation results from the splitting of one species into two, stabilizing selection is not usually thought to be involved in the initiation of speciation. Rather, stabilizing selection can contribute to speciation by maintaining phenotypic differences among populations that have evolved as a result of other forms of selection.

Disruptive Selection

Disruptive selection occurs when extreme phenotypes are favored over intermediate phenotypes within a population. This type of selection increases the variance of a trait and can divide a population into two distinct groups. If the two groups resulting from disruptive selection mate assortatively, then those groups can diverge to form species. Studies of Darwin’s finches on the Galápagos Islands by Peter R. Grant, B. Rosemary Grant, and others support the hypothesis that disruptive selection has contributed to speciation. This work has documented strong disruptive selection on beak size and shape caused by competition for similar food resources. Because of competition, selection favors a bimodal distribution of beak shapes as it facilitates the finches’ becoming specialized on different food resources. Differences in beak morphology are associated not only with variation in diet but also with variation in mating songs and mate choice. Therefore, it is thought that assortative mating is a by-product of disruptive selection on beak morphology in this system.

Directional Selection

Directional selection occurs when one extreme of the phenotypic spectrum is favored while the other extreme is disfavored within a population. The result of directional selection is that the mean phenotype of the population shifts in the direction of the favored phenotype.

Directional selection is most commonly thought to contribute to the formation of species when it operates differentially across habitats. For example, research conducted by David B. Lowry, Megan C. Hall, and John H. Willis found that coastal perennial and inland annual ecotypes of the yellow monkeyflower (Mimulus guttatus) are each adapted to their respective habitats in western North America. Inland plants have evolved early flowering and an annual life history to avoid reproducing during the hot summer seasonal drought. In contrast, coastal plants are sheltered from the drought by cooler temperatures and a persistent summer fog. Because these coastal plants have access to water year-round, they have evolved a perennial life history in which later flowering is favored. However, coastal plants must cope with oceanic salt spray, to which they have evolved salt tolerance. The flowering time and salt spray adaptations that differentiate the coastal perennial and inland annual ecotypes result in strong reproductive isolation between the ecotypes. In other words, adaptation to one of the two environments makes it difficult for individuals to survive and reproduce in the other environment, thus reducing gene flow between the ecotypes.

Uniform Directional Selection

The example of coastal perennial and inland annual ecotypes of M. guttatus illustrates how contrasting directional selection across habitats can contribute to reproductive isolation. However, different populations that experience uniformly acting directional selection could develop reproductive isolating barriers as a by-product of the same types of adaptations, because different populations might respond to the same selective regime through different types of genetic changes. If those genetic changes result in hybrid incompatibilities, then uniformly acting directional selection would be an underlying cause of reproductive isolation between the populations.

Forms of Selection in a Geographic Context

The classification of speciation into geographic categories (see chapter VI.3) is a useful framework for beginning to understand how natural selection can influence the evolution of reproductive isolating barriers. Allopatric speciation occurs when populations are completely geographically isolated, such that no migration occurs. If populations are separated long enough, the genomes of those populations can diverge to the point at which they remain distinct even if they come into secondary contact in the future. Under allopatric conditions, alleles contributing to reproductive isolation can spread through neutral genetic drift, sexual selection, and natural selection. Directional selection is likely the major form of selection that drives the evolution of reproductive isolating barriers between allopatric populations. Alleles under natural selection or genetically linked to genes under natural selection tend to have a higher substitution rate 4Nsμ (for which N is the population size, s is the coefficient of selection, and μ is the per generation per gene mutation rate) than neutrally evolving genes, for which the substitution rate is simply μ. Thus, adaptive mutations can be substituted at a faster rate than neutral mutations.

When diverging populations experience migration, successful differentiation is dependent on the strength of selection and the rate of gene flow between those populations. Migration between populations resulting in gene flow allows recombination to homogenize differences. It is therefore unlikely that uniform directional selection could result in the evolution of reproductive isolation between populations that exchange genetic material, because any universally adaptive mutation arising in one population would quickly be spread through all populations that exchange migrants. In contrast, divergent directional selection across habitats can result in reproductive isolation between populations even in the face of considerable migration. If selection is strong enough, regions of the genome that are involved in adaptations to different habitats will remain divergent. Conversely, alleles at neutral loci and alleles beneficial in both habitats will move between populations through migration, thus resulting in the homogenization of those regions of the genome.

Sympatric populations and those that exchange large numbers of migrants are more likely to require disruptive selection to evolve reproductive isolation. Theoreticians have explored this scenario extensively in an effort to understand how speciation might occur in sympatry, despite gene flow. However, efforts to find examples of sympatric speciation in nature have yielded only a few compelling examples (see chapter VI.3).

2. TYPES OF REPRODUCTIVE BARRIERS AND THE EFFECT OF SELECTION ON THEIR EVOLUTION

Considering that there are many agents of selection in nature, it is not surprising that natural selection can contribute to the process of speciation in multiple ways simultaneously. As pointed out by Dobzhansky and Clausen, speciation most often involves multiple reproductive isolating barriers driven by a combination of evolutionary forces. Types of barriers are listed in chapter VI.1. The contribution of natural selection to the evolution of those barriers is explained here.

Habitat Isolation

Habitat isolation is a barrier that reduces gene flow owing to adaptations of populations to divergent habitats. When populations become adapted to different habitats, individuals have a greater chance of surviving and mating in their native habitats compared with foreign ones. The reduced viability of immigrant individuals in the foreign habitat leads to a reduced rate of mating between native and foreign individuals. A classic example of habitat isolation occurs in Timena cristinae walkingstick insects. In California, populations of walkingsticks have become locally adapted to living on either Adenostoma fasciculatum plants or Ceanothus spinosus plants. Each of the walkingstick ecotypes has evolved cryptic coloration and morphological differences to blend in with the foliage of their respective host plants, presumably in response to visual predation by birds and lizards. Because of these morphological adaptations, walkingsticks have greater fitness when occurring on their native plant than on the foreign plant. This differential survival leads to more mating among walkingsticks adapted to the same plant species than among individuals adapted to the alternate plant species.

Temporal Isolation

Reproductive isolation that occurs as a result of populations mating at different times is called temporal isolation. Various forms of natural selection can lead to changes in the timing of mating and thereby cause reproductive isolation. Flowering-time divergence is commonly cited as a major form of temporal isolation in plants. Flowering-time evolution within plant species is frequently driven by selection imposed by abiotic stresses that cycle throughout the year. There are many examples of plants that have evolved adaptations to avoid flowering during annually recurring periods of environmental stress, such as drought or cold. These shifts in flowering as a result of habitat-mediated directional selection can lead to temporal reproductive isolation between populations adapted to different habitats.

Pollinator Isolation

Pollinator isolation is reproductive isolation resulting from pollinator behavior, such as preference for phenotypically different flowers. Pollinators impose natural selection on flowering plants that depend on them to mate. Shifts between pollinator guilds, such as from bees to hummingbirds, can lead to very strong reproductive isolation between plant taxa. For example, natural selection by different pollinator communities has led to morphological, color, and nectar production differences between bee-pollinated Mimulus lewisii and hummingbird-pollinated M. cardinalis. As a result, pollen is rarely transferred between these species.

Behavioral Isolation

Adaptation to different habitats can involve behavioral changes, which in turn cause changes in mating preference. For example, there is strong behavioral reproductive isolation between ecotypes of the apple maggot fly (Rhagoletis pomonella). Different ecotypes prefer either apple or hawthorn host plants. Behavioral experiments have demonstrated that flies show a strong preference for fruits of their respective host plant and will alter their orientation to go toward the desired odor associated with that plant. This leads to a reduction in mating between the two ecotypes inhabiting different trees and, consequently, assortative mating among individuals of the same ecotype. Behavioral isolation can also result from reinforcement, which we discuss later in this chapter.

Extrinsic Postzygotic Isolation

Extrinsic postzygotic isolation results when hybrids have reduced fitness because of maladaptation to either of the niches of the parental types. Extrinsic postzygotic isolation appears to be strong in a number of systems, including sticklebacks, leaf beetles, and big sagebrush. Stickleback (Gasterosteus) fish have evolved different ecotypes in shallow open water (limnetic) and deep (benthic) portions of lakes along the Pacific coast of Canada. Hybrids between benthic and limnetic ecotypes of sticklebacks have characteristics that are ecologically intermediate between the two ecotypes and as a result are maladapted to either of the lake habitats.

Intrinsic Postzygotic Isolation

Intrinsic postzygotic isolation manifests as hybrid lethality, inviability, or sterility and results from genic incompatibilities or chromosomal rearrangements (see chapter VI.8). Alleles contributing to intrinsic genetic incompatibilities can be spread by natural selection if an adaptive mutation directly contributes to an incompatibility or if an incompatibility allele is genetically linked to an adaptive mutation. Indeed, most of the identified genes underlying intrinsic postzygotic isolation appear to show molecular signatures of natural selection (see chapter VI.8). However, very little is known about the mechanisms of selection in these cases. In principle, external ecological factors could lead to the evolution of intrinsic postzygotic isolation.

There is now evidence that adaptation to internal genomic conflict can also drive the spread of isolating barriers. For example, Nitin Phadnis and H. Allen Orr recently showed that the gene Overdrive is a selfish genetic element that distorts Mendelian ratios in the gametes of F1 hybrids of two subspecies of Drosophila pseudoobscura, such that one allele has a higher probability of being in offspring than alternative alleles. Overdrive also causes hybrid sterility and thus contributes to reproductive isolation. Selfish genetic elements, such as Overdrive, can act as agents of selection on host genomes, which evolve mechanisms to repress them. The result is a coevolutionary battle in which the selfish element and the host genome both evolve in response to natural selection imposed by the other, increasing reproductive isolation in the process.

The Importance of Natural Selection in the Evolution of Reproductive Isolating Barriers

One way to assess the importance of natural selection in speciation is to compare the relative strengths of different types of reproductive isolating barriers between pairs of species. Some reproductive barriers, such as immigrant inviability and pollinator isolation, are thought to result primarily from natural selection. Other barriers, such as intrinsic reproductive isolation, could be the result of either neutral genetic drift or selection.

The most comprehensive analysis quantifying the strengths of multiple reproductive isolating barriers was carried out by Justin Ramsey, Douglas W. Schemske, H. D. Toby Bradshaw, and others between two species of Mimulus, M. cardinalis and M. lewisii. To date, the strengths of eight reproductive isolating barriers have been measured in this system. When combined, these barriers have a total strength of 0.9999, where 1.0 is complete reproductive isolation. These two species encounter each other infrequently in nature because M. cardinalis is adapted to lower elevations, while M. lewisii is adapted to higher elevations. As mentioned earlier, different pollinator communities visit these two species, limiting their pollen exchange. Thus, two ecological reproductive isolating barriers—habitat and pollinator isolation—are responsible for near-complete reproductive isolation in this system. Barriers that are less likely to be the result of natural selection, such as gametic isolation and F1 sterility, are weak between M. cardinalis and M. lewisii.

Some of the best evidence for a role of natural selection in speciation comes from studies that find the same reproductive isolating barriers evolving independently under similar ecological conditions. This repeated evolution of reproductive isolating barriers has been named parallel speciation. The walkingsticks provide an excellent example of parallel evolution of reproductive isolation. Studies of Timena cristinae walkingsticks by Patrik Nosil, Bernard J. Crespi, and others have identified eight geographically separated locations that contain Ceanothus- and Adenostoma-adapted ecotypes. Using molecular data, the researchers showed that these divergent reproductively isolated ecotypes have evolved independently. Across T. cristinae, individuals show preference for mating with other individuals adapted to the same host plant.

Studies of reproductive isolating barriers in a single system can provide insights into how speciation can occur. However, comparisons in multiple species pairs are necessary to draw broader conclusions about the relative importance of natural selection in the formation of reproductive isolation. Lowry and others (2008) recently compiled the strengths of multiple reproductive isolating barriers for 19 pairs of plant species. This comparative study revealed that prezygotic barriers were on average twice as strong as postzygotic barriers, with ecological barriers alone often accounting for near-complete reproductive isolation (see chapter VI.1). This is likely an underestimate of the importance of natural selection in the formation of reproductive isolation, because intrinsic postzygotic isolation can also be driven by natural selection. Overall, there is increasing evidence that natural selection accounts for the majority of reproductive isolating barriers involved in the process of speciation.

3. CONSIDERATIONS WHEN STUDYING NATURAL SELECTION AND SPECIATION

Because the process of speciation can take a long time, evolutionary biologists are restricted to studying a snapshot of the process for a given evolving pair of species. This limitation raises a number of important considerations when studying the role of natural selection in speciation.

First, the historical importance of particular reproductive isolating barriers cannot easily be determined. It is very difficult, if not impossible, to determine the historical order by which different reproductive isolating barriers accumulate and thus the role that natural selection plays over the entire process of speciation. The reproductive isolating barriers currently keeping species distinct might not have been the ones critical during the initial stages of speciation.

Another very important point to keep in mind when studying speciation is that reproductive isolation driven and maintained by ecological natural selection can collapse when environmental conditions change. A classic example of such species collapse involves cichlid fish in Lake Victoria, Africa—species that occur at different depths. The gradient of light quality correlated with depth has driven evolutionary changes in a light-absorbing opsin gene. It is thought that the local adaptation of female light perception to light at different depths has in turn led to changes in their preference for male coloration. Males of fish in shallow water, where blue wavelengths of light are abundant, have evolved blue coloration. Males occurring in deeper water, where available light is shifted to the red end of the visible light spectrum, are colored red. Under normal clear-water conditions, assortative mating occurs within deep and shallow populations because female cichlids prefer males that have the most visible colors. The recent introduction of anthropogenic pollution to the lake changed the way light passes to different water depths. This altered visibility has led to interbreeding followed by the collapse of closely related species. Eric B. Taylor, Janette W. Boughman, and others recently documented a similar collapse of deep- and shallow-water stickleback ecotypes in Enos Lake, British Columbia. In that case, changes to the lake caused by the introduction of a foreign crayfish apparently led to a breakdown of assortative mating. Thus, while ecological barriers are thought to evolve quickly during the process of speciation, they may not be sufficient to maintain species boundaries into the future.

Finally, it is challenging to infer whether divergent populations and ecotypes will ever become distinct species. It is clear from decades of research that not all adaptations within a species are necessarily involved in speciation. A major goal of speciation research is to determine why some adaptations are important to the process of speciation while other adaptations segregate within species without leading to appreciable reproductive isolation.

4. REINFORCEMENT

Reinforcement is the process whereby reproductive isolation increases as a direct response to natural selection against maladapted hybrids. This process is generally thought to involve three successive steps. First, two populations diverge in allopatry and accumulate partial reproductive isolation. Second, the two divergent populations come into secondary contact such that their ranges overlap partially or completely. Under this scenario, premating reproductive isolation is not complete and the two populations interbreed. There is selection against the interbreeding either because of direct costs of mating (e.g., copulatory damage) or, more commonly, because the hybrids have reduced fecundity, survival, or mating success. This selection against the production of hybrid offspring favors the evolution of greater reproductive isolation between the two diverging groups. Thus, the third step in reinforcement is the evolution of reproductive isolating barriers that reduce interbreeding or the production of maladapted hybrids.

Originally, reinforcement was thought to occur only when postzygotic isolation is very strong. However, recent theoretical models have shown that reinforcement can increase reproductive isolation between sympatric groups that are just beginning to diverge. During the process of reinforcement, natural selection against costly hybridization acts as a selective mechanism to drive the evolution of traits that prevent mating. The increase in reproductive isolation generally involves premating isolating barriers that either increase assortative mating or vary mate choice in a manner that decreases hybridization.

Reinforcement was originally termed the “Wallace effect” after Alfred Russel Wallace, who first articulated the idea that selection against hybrids might favor the evolution of reproductive isolation. Later, Dobzhansky described the process of reinforcement, as we understand it today, and stressed its potential importance in species formation. However, for much of the twentieth century, reinforcement was controversial.

Recombination and Reinforcement

Genetic recombination is at the core of the controversy regarding the occurrence of reinforcement in nature, because recombination between loci underlying the new prezygotic isolating barrier and loci involved in the original reproductive isolating barrier can lead to increased hybridization. Such increases in hybridization will ultimately result in the extinction or the homogenization of diverging populations.

The problem of recombination is best described using a hypothetical example involving two divergent ecotypes, red and blue. When these ecotypes interbreed, the hybrids produced are partially sterile, so selection favors decreased interbreeding between the ecotypes. A novel mutation then arises in the red ecotype, causing strong preference for red individuals over blue individuals. Similarly, a new mutation arises at the same locus causing the blue ecotype to prefer blue individuals. This type of reinforcement is termed a two-allele mechanism because two alternative alleles at a single locus result in assortative mating. Both alleles prevent the formation of partially sterile hybrids and are thus favored. The problem with recombination arises when the two ecotypes hybridize. Recombination between the preference locus and loci contributing to sterility in hybrids leads to the production of individuals that mate with the opposite ecotype. As a result, assortative mating can break down, and the ecotypes may fuse into a single population or go extinct.

Despite the problem that recombination can pose for reinforcement, recently developed theoretical models have convincingly demonstrated that selection can overcome the homogenizing effect of recombination and maintain assortative mating under a variety of scenarios. Furthermore, empirical investigations have found evidence for reinforcement in insects, amphibians, birds, mammals, and plants, thus indicating the potential importance of this process in speciation across a wide range of biological diversity.

One way the problem of recombination can be alleviated is for the same allele to increase assortative mating in both diverging taxa. Joseph Felsenstein first discussed this one-allele mechanism of reinforcement in 1981. Under this scenario, a novel allele is favored in both ecotypes because it decreases hybridization in either genetic background. This is a one-allele mechanism because the same allele reduces hybridization in both diverging ecotypes. Empiricists were skeptical that this sort of mechanism could actually exist in nature until Daniel Ortiz-Barrientos and Mohamed Noor found strong evidence for a one-allele mechanism of reinforcement in sympatric populations of Drosophila pseudoobscura and D. persimilis. A single allele causes both the D. pseudoobscura females to mate more with D. pseudoobscura males, and the D. persimilis females to mate more with D. persimilis males.

Geographic Patterns of Reinforcement

Instances of reinforcement are often identified when selection causes increased reproductive isolation in the sympatric but not the allopatric parts of the range of two diverging groups. The pattern arises because reinforcing selection occurs only when two diverging groups interact and attempt to mate in overlapping areas of their ranges. Traits favored by reinforcement in sympatry will often be neutral or even disfavored in allopatry and therefore not spread into those areas of the range.

A classic example of phenotypic divergence resulting from reinforcement is flower color variation in species of the genus Phlox. Phlox drummondii and P. cuspidata both have the same light-blue flower color in allopatric regions of their range. In sympatry, P. drummondii has dark-red flower color, while P. cuspidata retains a light-blue color. Robin Hopkins, Mark D. Rausher, and Donald A. Levin have shown that this flower color change reduces the production of hybrids between these species. Although both species of Phlox are pollinated by the same species of butterflies, individual pollinators rarely move pollen between the Phlox species if they have different flower colors. Dark-red-flowered P. drummondii plants exchange pollen with P. cuspidata plants less frequently than light-blue-flowered P. drummondii plants exchange pollen with P. cuspidata plants. Thus, trait evolution within a species can result in increased reproductive isolation between species.

The first 75 years of research on reinforcement was predominantly concerned with determining whether the process could exist. Now that there are conclusive theoretical and empirical studies supporting its occurrence, research has turned its focus toward determining how, why, and when it occurs. Compared with many areas of speciation research, which rely heavily on empirical studies, theoretical work has shaped our understanding of the process of reinforcement. Of the examples of reinforcement identified in nature, its genetic basis has been identified only in the Phlox system, where mutations in two genes involved in the production of anthocyanin floral pigments are responsible for a change in flower color. More work is required to determine the patterns, if any, in the types of mutations; the genetic architecture of traits underlying reinforcement; and the strength of selection acting during the process of reinforcement.

5. FUTURE DIRECTIONS

Many questions remain unresolved regarding natural selection and speciation. While major progress has been made recently in quantifying reproductive isolating barriers, we still have a poor understanding of the role that different barriers play over the entire process of speciation. An important question is, Are the reproductive isolating barriers important during the initiation of speciation different from those that prevent the collapse of species pairs later in the process?

Major progress will be made through the identification of genes involved in reproductive isolation and the mechanisms that caused their spread. Many of the genes discovered to date that underlie reproductive isolation appear to be driven by natural selection (see chapter VI.8). However, the mechanisms by which natural selection has driven the spread of reproductive isolating alleles at those genes are for the most part unknown. Further, most genes identified as underlying reproductive isolation are involved in intrinsic postzygotic isolation. Greater focus should be placed on identifying the genes underlying other types of reproductive isolating barriers that more clearly involve ecological adaptations. Once those genes are identified, follow-up studies should be conducted to determine the geographic distribution of alleles at those genes to better understand how natural selection spreads reproductive isolation during the process of speciation.

See also chapter III.8, chapter VI.2, and chapter VI.7.