Evolution, 4th Edition Chapter 9-17 PDF

 9 Species and Speciation In the Rift Valley of eastern Africa, just south of the equator, lie three great lakes and many smaller ones. Lakes Tanganyika and Malawi are deep and old, having been f...

 9 Species and Speciation In the Rift Valley of eastern Africa, just south of the equator, lie three great lakes and many smaller ones. Lakes Tanganyika and Malawi are deep and old, having been formed by the separation (rifting) of two continental plates. Lake Victoria, in contrast, is broad and shallow, lying in a basin that was dry only 15,000 years ago. These lakes harbor a few species of catfishes, spiny eels, and other fish families, but more than 90 percent of all the fish species are cichlids, a family that includes species well known to tropical fish hobbyists. Lake Tanganyika has at least 250 species of cichlids, Lake Victoria between 450 and 530 species, and Lake Malawi at least 480 species. (The American Great Lakes, in comparison, have only about 175 species of fishes, of all kinds.) These cichlid fishes are extraordinarily diverse in coloration, form, feeding habits, and habitat use (FIGURE 9.1). Different species eat insects, snails, detri- tus, rock-encrusting algae, aquatic plants, phytoplankton, zooplankton, baby fishes, and larger fishes. Some species are specialized to feed on the scales of other fishes, and one has the gruesome habit of plucking out other fishes’ eyes. The teeth of some closely related species differ more than do those of some whole families of fishes. Many of these habits and morphologies have evolved convergently in the different lakes. Phylogenetic analyses show that the 250 cichlids in Lake Tanganyika have evolved from at most 16 original species. The cichlids of Lake Victoria have multiplied faster than any other group of vertebrates on Earth: the 450-plus species evolved from just 5 origi- nal ancestral species in perhaps only 15,000 years [95, 105]. A male gray tree frog inflates his vocal sac as he calls to attract females. Female frogs re- spond almost exclusively to their own species’ calls, which are a barrier to interbreeding. Male calls differ between two morphologically indistinguishable species of gray tree frogs in eastern North America. Hyla chrysoscelis has 12 pairs of chromosomes, whereas H. versicolor is a tetraploid, with 24 pairs. 214   CHAPTER 9 FIGURE 9.1 Examples of the diversity of (A) cichlid fishes in Lakes Tanganyika (at left) and Malawi (at right). Ecologically and morpho- Petrochromis Petrotilapia logically similar forms have evolved inde- pendently in both lakes. (A) Rock-dwelling species with rasping jaws. (B) Open-water fish-eaters. (C) Fleshy-lipped species that (B) suck prey from crevices. (D) Rock-dwellers. (E) Hump-headed species. (F) Slender, Bathybates Rhamphochromis striped species. (From ). (C) Lobochilotes Placidochromis (D) Tropheus Pseudotropheus (E) Cyphotilapia Cyrtocara (F) Julidochromis Melanochromis What caused this explosion of diversity? Do the number and ecological variety of species depend only on current ecological conditions, such as how many differ- ent kinds of resources can sustain different species? Or do they reflect the rate at which new species have arisen? Why should the rate of speciation have been so high in this fish family, and only in these lakes? Has speciation been caused by the fishes’ mating patterns? By sexual selection on coloration? By adaptation to differ- ent ecological niches? By genetic drift? Fundamentally, what we want to know is: How do new species form? Darwin first came to believe in evolution when he realized that different islands in the Galápagos archipelago harbor different forms of mockingbirds and a variety of similar finches. That these forms were similar, yet subtly different, could most plausibly be explained by supposing that they had descended, with slight modifica- tions, from a common ancestor. Pursuing this reasoning, Darwin concluded that all species of birds—indeed all species of animals, and finally all living things—may have originated by successive branching of lineages throughout the history of life, from a single common ancestor. Modern research has affirmed that this is indeed how the Futuyma enormous Kirkpatrick diversity Evolution, 4e of organisms arose. The forks in the great Tree of Life were caused by speciation, the process by which one species gives rise to two. Sinauer Associates Troutt Visual Services Evolution4e_0901.ai Date 11-02-2016  Species and Speciation   215 (A) (B) (C) FIGURE 9.2 Can you distinguish the species? (A, B) Gray and rufous morphs of the east- ern screech owl (Megascops asio). (C) The western screech owl (Megascops kennicottii). What Are Species? Several definitions of “species”—which is Latin for “kind”—are used by biolo- gists. It is important to bear in mind that a definition is not true or false, because the definition of a word is a convention. Probably no single definition of “species” suffices for all the contexts in which a species-like concept is used. For Linnaeus and other early taxonomists, species were simply groups of organ- isms that could be distinguished. But as knowledge of organisms grew, this crite- rion became inadequate. For example, two kinds of small owls in eastern North America look very different: one is gray and the other bright reddish brown (FIG- URE 9.2A,B). Nevertheless, they are clearly the same species: the two forms sound the same, they interbreed, and a brood may include both color forms—which are a simple one-locus polymorphism (with rufous dominant over gray). But the gray form of this species, the eastern screech owl (Megascops asio), is almost indis- tinguishable in appearance from another owl that has a very different voice and that is recognized as a distinct species—the western screech owl (M. kennicottii; FIGURE 9.2C). The two species can be completely distinguished by mitochondrial DNA , indicating that even though they coexist in Texas, there is little or no gene flow between them. They are separate gene pools. Cases such as the screech owls led to the concept of species as groups of indi- viduals that interbreed. Ernst Mayr formalized this idea in what he called the biological species concept (BSC), defined as follows: “Species are groups of actually or potentially interbreeding populations, which are reproductively isolated from other such groups.” Reproductive isolation means that any of several biological differences between the groups greatly reduce gene exchange between them, even if they are not geographically separated. The BSC does not require that species be 100 percent reproductively isolated—there can be a little genetic “leakage” between species through hybridization. Although genetic and phenotypic differences do not define species according to the BSC, those differences enable us to recognize and distin- guish them. Note that an inability to form hybrid offspring, or sterility of hybrids, is not a necessary criterion of species: it is only one of many ways in which gene exchange may be reduced or prevented. The biological species concept was developed partly to acknowledge variation, both within a single population (such as the color morphs of the eastern screech owl) and among different geographic populations, which often show evidence of Futuyma Kirkpatrick Evolution, 4e Sinauer Associates Troutt Visual Services Evolution4e_0902.ai Date 11-02-2016  216   CHAPTER 9 interbreeding where they meet. The BSC also rec- ognizes cases of “sibling species” (such as the gray forms of the two screech owls), which are almost identical in appearance and are often discovered by differences in ecology, behavior, chromosomes, or genetic markers. The discovery that the European mosquito Anopheles maculipennis is actually a cluster of nine sibling species had great practical importance because some transmit human malaria and oth- ers do not [3, 39]. The term “sibling species” differs from sister species, which are two species descended from a single ancestral species, and are therefore one another’s closest relatives. The biological species concept is the most widely used definition among biologists, and it can be applied to the majority of sexually reproducing spe- cies alive on Earth. It does, though, have limitations. Reproductive isolation evolves gradually, as we will see. So interbreeding versus reproductive isolation is not an either/or, all-or-none distinction. Neverthe- less, there are countless examples of closely related forms that occur in the same area, can be distin- guished by genetic and phenotypic differences, and interbreed very little or not at all. They are unequivo- cally distinct, real species. Pygmy nuthatch Brown-headed nuthatch The greatest practical limitation of the BSC is in FIGURE 9.3 The geographic ranges of the pygmy nuthatch (Sitta determining whether populations that are geographi- pygmaea, left), in western North America, and of the brown-headed cally separated (allopatric) belong to the same species nuthatch (Sitta pusilla, right), in the southeastern United States, are (FIGURE 9.3). The BSC requires that we make a judg- separated by hundreds of miles in which neither bird occurs. They ment call as to whether they would interbreed if they differ in voice and subtly in color pattern. It is difficult to tell if they came into contact under natural conditions. Climate are different biological species. change in the past and human changes to the envi- ronment at present have brought formerly isolated populations together. In some such cases, the popula- tions remained distinct, but in other cases they interbred, showing that they were not fully distinct species. One could test for reproductive isolation experimentally, for example in the lab or garden, but this is impractical or even impossible to do with many species (e.g., giant squids). Moreover, some species that mate under artificial conditions will not do so in nature, and hybrid offspring that are viable and fertile in the lab may not survive in nature. In practice, deciding whether geographically isolated populations are species is at times somewhat arbitrary. Commonly, allo- patric populations have been classified as species if their differences in phenotype or in DNA sequence are as great as those usually displayed by species in the same group that are sympatric (in the same location). A similar approach is taken with classifying fossils into species, since paleontologists cannot study the mating behavior or hybrid survival of extinct ammonites or dinosaurs. Another limitation of the BSC is that it does not apply to organisms that do not reproduce sexually. Bacteria pose particular challenges. Although they do not have meiotic sex, they do exchange genetic material in other ways. Species of bacteria, such as Escherichia coli and Salmonella typhimurium, were traditionally recognized by differences in their metabolic capabilities. More recently, genetic similarity has been used to group individuals into species. Although bacteria can acquire new genes from even distantly related organisms, most homologous recombination (“sex”) occurs within traditionally recognized species. Futuyma Kirkpatrick Evolution, 4e Sinauer Associates Troutt Visual Services  Species and Speciation   217 (A) 0.99 FIGURE 9.4 The eastern European fire- 0.9 bellied toad (Bombina bombina) and the Allele frequency western European yellow-bellied toad (B. B. bombina variegata) meet and interbreed in a narrow 0.5 hybrid zone. The two species differ in loci that code for enzymes and several morpho- logical features. (A) Average allele frequen- 0.1 cy at six enzyme loci. (B) A morphological score based on seven characters. Red and blue dots represent two different 60-km 0.01 transects in Poland. The clines in enzyme –30 –20 –10 0 10 20 30 40 loci and morphological features are coinci- Distance from center of cline (km) dent, suggesting that this hybrid zone was formed by contact between two formerly (B) 0.99 B. variegata allopatric populations. (After.) Morphological score 0.9 0.5 0.1 0.01 –30 –20 –10 0 10 20 30 40 Distance from center of cline (km) These and other considerations have inspired several alternative species defi- nitions. Some systematists prefer the phylogenetic species concept (PSC), which emphasizes species as the outcome of evolution—the products of a history of evolutionary divergence. In one widely accepted definition, lineages are different species if they can be distinguished: a phylogenetic species is an irreducible (basal) cluster of organisms diagnosably different from other such clusters, and within which there is a parental pattern of ancestry and descent. The phylogenetic and biological species concepts have different uses, and tend to be used by different groups of researchers. The PSC can be useful for classifica- tion, because unlike the BSC, it can be applied to allopatric populations, such as those on different islands, in which reproductive isolation is difficult or impossible to assess. Although some systematists use the PSC in classifying organisms, most evolutionary biologists use one or another variant of the BSC, because they view the evolution of reproductive isolation as the key event that enables sexually repro- ducing lineages to evolve independently and generate biological diversity. Without the evolution of reproductive isolation, there would be only one (or at most a few spatially separated) species of cichlid in each of those African lakes. No matter which species concept is adopted, some populations of organisms cannot be unambiguously assigned to one species or another, because the features that distinguish species (by any definition) evolve gradually. There exist graded levels of gene exchange among adjacent (parapatric) populations and sometimes between more or less distinct populations that are sympatric. Species as recog- nized by the BSC are ambiguous in hybrid zones, which exist where genetically Futuyma Kirkpatrick Evolution, 4e distinct populations meet and interbreed to a limited extent, but in which there Sinauer Associates exist partial Troutt Visualbarriers Services to gene exchange (FIGURE 9.4). Hybridization occurs, at least Evolution4e_09.04.ai Date 01-23-2017 218   CHAPTER 9 FIGURE 9.5 Advantageous alleles have H. numata spread by introgression between distantly re- lated species of Heliconius butterflies in South H. ethilla America. The phylogeny is based on many genes. The DNA sequence of two genes that H. hecale control color pattern shows that H. timareta ssp. nov. acquired the “postman” pattern in H. pardalinus sergestus the hindwing from H. melpomene amaryl- lis, and that H. elevatus acquired the “rayed” H. elevatus hindwing pattern from H. melpomene aga- lope/malleti. (From ; large wing images H. pardalinus butleri courtesy of J. Mallet.) H. timareta florencia H. timareta ssp. nov. H. heurippa H. cydno H. melpomene rosina H. melpomene melpomene H. melpomene amaryllis H. melpomene aglaope/malleti Rayed Postman Heliconius hindwing patterns occasionally, among sympatric species in many groups of plants and animals , and genes are sometimes incorporated into the gene pool of one species from another, a process called introgression (or introgressive hybridization). Some such genes may enhance adaptation. For instance, Heliconius butterflies are distaste- ful to predators and have warning coloration: predators do not attack butterflies with this pattern after one or two experiences in which they learn to associate the coloration with distastefulness. Alleles that determine part of the color pattern of the wings of certain Heliconius species have spread among even distantly related species (FIGURE 9.5). Biological species are seldom distinguished in practice by directly testing their propensity to interbreed or their ability to produce fertile offspring. Indeed, this is usually not necessary. Morphological and other phenotypic characters are the usual evidence used for diagnosing sympatric species (FIGURE 9.6), because they can serve as markers that indicate reduced gene flow—that is, reproductive isola- tion—among sympatric populations. If a sample of sympatric organisms falls into two discrete clusters that differ in multiple characters, it is likely to represent two species. In modern studies, genetic markers are often used to reveal the existence of two or more sympatric species. A polymorphic locus that shows few heterozy- gotes, and so departs strongly from Hardy-Weinberg equilibrium, is a signal that there are likely to be more than one species. (BOX 9A provides an example.) Futuyma Kirkpatrick Evolution, 4e Sinauer Associates Troutt Visual Services  Species and Speciation   219 Phrynosoma FIGURE 9.6 An example of species distinguished by morphological charac- ters. These seven species of horned lizards (Phrynosoma) from western North America cornutum can be distinguished by differences in solare the number, size, and arrangement of modestum horns and scales as well as body size and proportions, color pattern, and habitat. Good scientific drawings can often show detailed features better than photographs can, especially when the critical features coronatum are subtle. (From.) m’calli platyrhinos douglassi BOX 9A Diagnosis of a New Species Each species in the leaf beetle genus Ophraella feeds on one species or a few related species of plants. O. notu- lata, for example, has been found feeding only on two species of Iva along the East Coast of the United States. This species is most readily distinguished from other species of Ophraella by the number and pattern of dark stripes on each wing cover. Some leaf beetles found in Florida closely resembled O. notulata but were collected on ragweed, Ambrosia ar- temisiifolia. This host association suggested the possibility that these beetles were a different species. In a broader study of the genus, one of the authors of this book (DJF) collected samples of beetles from both Ambrosia and Ophraella slobodkini Iva throughout Florida and examined them by enzyme electrophoresis. He found consistent differences in morphological characters, such as the shape of one of the allele frequencies between samples from Iva and from mouthparts and the relative length of the legs. Later stud- Futuyma Kirkpatrick Evolution, 4e SinauerAmbrosia Associates at three loci, even in samples from both plants ies showed that adults and newly hatched larvae strongly in theServices Troutt Visual same locality. In the most extreme case, one allele prefer their natural host plant (Ambrosia or Iva) when Evolution4e_0906.ai Date 11-02-2016 had an overall frequency of 0.968 in Ambrosia-derived given a choice, and that the beetles mate preferentially specimens, but was absent in Iva-derived specimens, with their own species. In laboratory crosses, viable eggs in which a different allele had a frequency of 0.989. No were obtained by mating female Ambrosia beetles with specimens had heterozygous allele profiles that would males from Iva, but not the reverse. Few of the hybrid suggest hybridization. Later study showed differences in larvae survived to adulthood, and none laid viable eggs. mitochondrial DNA as well. Thus these genetic markers Based on all of this evidence, the author concluded that were evidence of two reproductively isolated gene pools. the Ambrosia-associated form is a distinct species, and he A careful examination then revealed average differences named it Ophraella slobodkini in honor of the ecologist between Ambrosia- and Iva-associated beetles in a few Lawrence Slobodkin. 220   CHAPTER 9 (A) M. lewisii (B) M. cardinalis FIGURE 9.7 Pollinator isolation in monkeyflowers. (A) Mimulus lewisii has the broadly splayed petals characteristic of many bee- pollinated flowers. (B) M. cardinalis has the red coloration and narrow, tubular form that have evolved independently in many bird-pollinated flowers. (C) Some F2 hybrids, showing the varia- tion that Schemske and Bradshaw used to analyze the genetic basis of differences between these two species. (From.) (C) Reproductive Isolation Gene flow between biological species is prevented by biological differences called reproductive isolating barriers (RIBs), also referred to as isolating mechanisms. Under the biological species concept, speciation is the evolution of biological bar- riers to gene flow, and so understanding the evolution of reproductive isolating barriers tells us how new species evolve. The total degree of reproductive isolation between two species may result from several RIBs that act in sequence—and some potential RIBs may not come into play. For example, the monkeyflower Mimulus lewisii is distributed in the Sierra Nevada of California at higher elevations than its close relative Mimulus cardinalis, although they both occur at intermediate elevations. Mimulus lewisii has pink flowers with a wide corolla and is pollinated by bees, whereas M. cardinalis has a narrow, red, tubular corolla and is pollinated by hummingbirds (FIGURE 9.7). Although almost no hybrids are found where the species occur together, the spe- cies can be readily crossed, and they produce viable, fertile hybrids. To understand the roles played by different isolating barriers between these spe- cies, Douglas Schemske, Toby Bradshaw, and their colleagues performed a massive field experiment. They bred a large number of F2 hybrids and planted them in an area where the two species coexist. The F2s have far greater phenotypic variation than the parental species, and they have novel combinations of traits. By amplify- ing the variation this way, the researchers were able to determine which of 12 floral traits that distinguish the parental species are important to reproductive isolation. They went further by using a quantitative trait loci (QTL) study (see Chapter 6) to reveal the genes underlying those traits. At least four traits affect the type of pollinator that is attracted to a flower, which in turn determines which individu- als exchange Futuyma Kirkpatrick genes. Evolution, 4e The difference between the species in some of these traits is Sinauer Associates based on as few as one to as many as six QTL, so a change to one or a few genes Troutt Visual Services can greatly Evolution4e_0907.ai Date affect reproductive isolation. 11-02-2016 The investigators were able to quantify the contribution that different mecha- nisms make to reproductive isolation, in sequence (FIGURE 9.8). Separation by ele- vation alone reduces gene exchange by 59 percent. Among plants living at the same  Species and Speciation   221 elevation, pollinator fidelity alone is 98 percent effective. If a flower receives both species’ pollen, the conspecific pollen (that Elevation Reproductive barrier is, pollen from the same species) fertilizes the ovules at least Pollinator 70 percent of the time. The germination of F1 hybrid seeds is reduced by 20 percent compared with nonhybrids, but a hybrid Pollen precedence M. cardinalis seed that does germinate is just as viable. Hybrids produce fewer seeds, however, and they produce much less viable pol- Hybrid germination len. But because isolation by elevation and pollinator behavior is so great, the later barriers—reduced production, viability, and Hybrid fertility fertility of hybrids—hardly come into play at all. The Mimulus species illustrate some of the many kinds of 0 0.25 0.5 RIBs (TABLE 9.1). Prezygotic barriers reduce the likelihood Contribution to isolation M. lewisii that hybrids are formed. These include such factors as separa- FIGURE 9.8 Relative contributions of successively acting isolat- tion of the species in different habitats, pollination by differ- ing mechanisms between the monkeyflowers Mimulus lewisii ent animals, mating at different seasons, mating preferentially and M. cardinalis. Elevational separation and pollinator isola- with conspecifics, and failure of gametes to unite even if mat- tion account for almost all the reproductive isolation. In places ing occurs. Postzygotic barriers reduce gene exchange between where both species occur, pollinators provided almost com- populations even if hybrid zygotes are produced. They consist plete reproductive isolation. (After ; photos courtesy of of reduced hybrid viability (survival) or reproduction (fertil- D. W. Schemske and H. D. Bradshaw, Jr.) ity). Both classes of barriers are often asymmetric: for example, females of species A may be less inclined to mate with males of species B than females of B are to mate with males of A , or F1 hybrids between the species may differ in viability, depending on the direction of the cross [104, 108]. Since prezygotic isolating mechanisms act before postzygotic mechanisms, they have a greater opportunity to restrict gene flow. A second reason why the distinc- tion between prezygotic and postzygotic mechanisms is useful is that different kinds of selection act on them, as we will see shortly. It is often difficult to tell which isolating barrier was the original cause of spe- ciation. A character difference that contributes to reproductive isolation now may have evolved partly in geographically segregated populations before they became TABLE 9.1 A classification of isolating barriers I. Premating barriers: features that impede transfer of gametes to members of other species A. Ecological isolation: potential mates do not meet 1. Temporal isolation: species breed at different seasons or times of day 2. Habitat isolation: species mate and breed in different habitats 3. Immigrants between divergent populations do not survive long enough to interbreed B. Potential mates meet but do not mate 1. Sexual isolation in animals: individuals prefer mating with members of their own species 2. Pollinator isolation in plants: pollinators do not transfer pollen between species II. Postmating prezygotic barriers: mating occurs, but zygotes are not formed A. Mechanical isolation: reproductive structures of the sexes do not fit B. Copulatory isolation: female is not stimulated by males of the other species C. Gametic isolation: failure of fertilization III. Postzygotic barriers: hybrids are formed but have reduced fitness A. Extrinsic: hybrids have low fitness for environmental reasons 1. Ecological inviability: hybrids are poorly adapted to both of the parental habitats 2. Behavioral sterility: hybrids are less successful in obtaining mates B. Intrinsic: low hybrid fitness is independent of environmental context 1. Hybrid inviability: reduced survival is due to genetic incompatibility 2. Hybrid sterility: reduced production of viable gamates Source: After , in part. Futuyma Kirkpatrick Evolution, 4e Sinauer Associates Troutt Visual Services Evolution4e_09.08.ai Date 11-18-2016 (A) (B) (C) (D) (E) (F) FIGURE 9.9 Prezygotic isolation takes many forms, illustrated on different species of plants. (D–F) Examples of sexual isola- by some species that have been extensively studied. (A–C) tion based on different sensory modalities. (D) Female Physalae- Three modes of prezygotic isolation. (A) Seasonal isolation: the mus frogs respond almost exclusively to the calls of conspecific band-rumped storm-petrel (Oceanodroma castro) includes two males. A calling male P. pustulosus is shown here. (E) In moths genetically different populations that mate at different times of year and many other animals, sexual isolation is based on different. (B) Temporal isolation: related species of periodical cicadas chemical signals. Two forms of the European corn borer (Ostrinia (Magicicada) have either 17- or 13-year life cycles, and rarely nubilalis) are strongly isolated by responses of males to different emerge in the same year. (C) Ecological isolation: closely re- female sex pheromones. (F) Males of Heliconius pachinus rec- lated species of ladybird beetles (Henosepilachna) feed and mate ognize conspecific females by their wing color pattern. different species, partly during the process of speciation, and partly after the repro- ductive barriers evolved. Because genetic differences continue to accumulate long after two species achieve complete reproductive isolation, some of the genes, and even some of the traits, that now confer reproductive isolation may not have been instrumental in forming the species in the first place. Such information can be obtained by studying populations that have achieved reproductive isolation only very recently. Prezygotic barriers In many plants and animals, prezygotic barriers are the most important isolat- ing mechanisms. There are many kinds of barriers, depending on the biology of the organism (FIGURE 9.9). Species may be temporally isolated by mating at dif- ferent times of year, or even in different years. Ecological isolation results when ecological differences, for example habitat preference, contribute to genetic bar- riers [64, 93]. For example, two Japanese species of herbivorous ladybird beetles (Henosepilachna) feed on different genera of host plants (Cirsium and Caulophyl- lum). Each species mates exclusively on its own host plant, and this ecological seg- regation appears to be the only barrier to gene exchange. Sexual isolation is an important barrier to gene flow among sympatric species of animals that fre- quently encounter each other but simply do not mate. Commonly, females will not respond to inappropriate male vocalizations or other display signals. Many birds, fishes, and jumping spiders are sexually isolated by visual signals. In many groups Futuyma Kirkpatrick Evolution, 4e Sinauer Associates of animals, sexual isolation is based on differences in sex pheromones. Troutt Visual Services Evolution4e_0909.ai Date 11-02-2016  Species and Speciation   223 (A) (B) FIGURE 9.10 Differences in genitalia can contribute to reproductive isolation between species if copulation between them occurs. (A) The genital arch in male Drosophila is involved in transferring sperm to females. Its shape differs among closely related species, as the close-ups show: (B) D. sechellia, (C) D. mauritiana, (D) D. simulans. This morphological fea- ture is almost the only one by which these 100 μm species differ. (A from ; B, C, and D courtesy of J. R. True.) (C) (D) Gametic isolation occurs when gametes of different species fail to unite. This barrier is important in many externally fertilizing species of marine invertebrates that release eggs and sperm into the water. Because cell surface proteins determine whether or not sperm can adhere to and penetrate an egg, divergence in these proteins can result in gametic isolation. Among species of abalones (large gas- tropods), the failure of heterospecific eggs and sperm to unite is related to the high rate of divergence in the amino acid sequences of both lysin (the sperm protein that dissolves the egg’s vitelline envelope) and the vitelline envelope protein with which it interacts (see Chapter 10). In cases that fall in between premating and postmating isolation, mating occurs but fertilization does not. In many groups of insects and some other taxa, the genitalia of related species differ in morphology. There is evidence that females terminate mating, and prevent transfer of sperm, if a male’s genitalia do not provide suitable tactile stimulation (FIGURE 9.10). Postzygotic barriers 0.08 Proportion attacked Postzygotic barriers consist of reduced survival or reproductive rates of hybrid 0.06 zygotes that would otherwise backcross to the parent populations. These barriers can be classified as either extrinsic or intrinsic, depending on whether or not their effect 0.04 depends on the environment. Intrinsic isolation is based on interactions between genes from two populations, and is often more permanent than extrinsic isolation. 0.02 Extrinsic postzygotic isolation is often based on reduced survival because of 0 ecological factors. In some cases, the parent species are adapted to different envi- melpomene F1 hybrid cydno ronments; the hybrid may be poorly adapted to both. A simple example is provided FIGURE 9.11 Model butterflies with the by hybrids between species of Heliconius butterflies that are distasteful to birds and color pattern of the F1 hybrid between have different patterns of warning coloration. Birds learn to associate common Heliconius melpomene and H. cydno color patterns with distastefulness, but are likely to attack butterflies with rare, were attacked by birds significantly more unfamiliar phenotypes, such as hybrids. Researchers placed artificial butterflies, frequently than those with the pattern of with wing patterns of two species and their F1 hybrid, in a tropical forest, and either parent. The low survival of hybrids is scored the number that were damaged by attacking birds. Those with hybrid an example of postzygotic isolation caused color patterns were more frequently attacked (FIGURE 9.11). by an extrinsic factor. (From.) Futuyma Kirkpatrick Evolution, 4e Sinauer Associates Troutt Visual Services Evolution4e_0910.ai Date 11-02-2016 224   CHAPTER 9 Ancestor Postzygotic isolation is intrinsic if hybrids suffer high mortality, or are partially A1A1B1B1 or entirely sterile, irrespective of environment. The causes of intrinsic postzy- gotic isolation and its genetic bases are diverse. Reduced hybrid viability is largely Geographic caused by incompatible interactions among genes from the two populations when separation they occur together in hybrids. Hybrid fertility may be reduced by incompatible Population 1 Population 2 genes or by differences in the number or structure of chromosomes. Bear in mind that the genetic differences that cause these effects may have evolved after prezy- A1A1B1B1 A1A1B1B1 gotic barriers, so we cannot assume that they were the cause of speciation. Incompatible interactions between genes inherited from the two parents were Time Genetic postulated by Theodosius Dobzhansky in 1937 and by Hermann Muller in divergence 1942 , and are often referred to as Dobzhansky-Muller incompatibilities (DMIs). The Dobzhansky-Muller hypothesis is clever because it explains how incompat- A2 A2B1B1 A1A1B2B2 ibilities between populations can originate without ever producing incompatibili- ties within a population (FIGURE 9.12). Imagine that the ancestor of the two spe- cies had genotype A1 A1B1B1. That species was then divided into two populations F1 hybrids A1A2B1B2 by a geographic barrier. In one population, allele A 2 spreads to fixation (perhaps because of adaptation to local conditions). This population is now A 2 A 2 B1B1. In the second population, allele B2 spreads to fixation, so this population becomes FIGURE 9.12 Dobzhansky-Muller in- A1 A1 B2 B2. During this period, alleles A 2 and B2 have never been in the same compatibilities (DMIs) can evolve when population, so there is no reason they should have been selected to function well geographically separated populations together. If they are incompatible, hybrids between the two populations will have become fixed for different alleles at two low fitness. loci. (After.) A simple example has been described for a cross between strains of the mouse- ear cress Arabidopsis thaliana from different regions. Both strains have two paral- ogous loci (call them α and β), formed by duplication. In one strain, the α locus is nonfunctional, but the β locus is functional. The other strain has a functional α but a nonfunctional β. The F1 offspring of a cross between the strains are viable, but in the F2 generation, some recombinant offspring are homozygous for nonfunctional alleles of both α and β genes—a lethal combination. DMIs between Drosophila simulans and D. mauritiana cause male F1 hybrids to be sterile, while females are fertile. The genetics of the hybrid male sterility have been studied with laboratory crosses that produce different combinations of chro- mosome segments. Two results emerge. The first is that many combinations of chromosomes from the two species reduce male fertility, showing that there are many DMIs throughout their genomes. The second is that male sterility is caused by interactions between the autosomes of simulans and the X chromosome of mau- ritiana. This reflects a general phenomenon called Haldane’s rule: hybrid sterility or hybrid inviability is often limited to the heterogametic sex. (The heterogametic sex is the one with two different sex chromosomes, while the homogametic sex has two sex chromosomes of the same type.) In mammals and most insects, males are XY and thus are the heterogametic sex. In birds and butterflies, the situation is reversed: females have two different kinds of sex chromosomes. Thus male hybrids are frequently sterile in mammals (for example, mules), while female hybrids are frequently sterile in birds. DMIs have many causes. Gene regulation can be anomalous due to a mismatch between cis- and trans-regulatory elements from the two species. Intragenomic conflict (see Chapter 12) appears to be a common cause (see below) [18, 75]. DMIs can also be manifestations of cytonuclear incompatibility. For example, hybrids between different geographic populations of a marine copepod have reduced sur- vival and fecundity if their mitochondria and nuclear genome come from different populations (FIGURE 9.13). Many sister species are distinguished by chromosome rearrangements: structural differences between the chromosomes (see Chapter 4). Two common Futuyma Kirkpatrick Evolution, 4e Sinauer Associates Troutt Visual Services  Species and Speciation   225 0.65 FIGURE 9.13 Crosses show that the low fitness of hybrids between populations of the cope- pod Tigriopus californicus is caused by a genetic 0.60 mismatch between mitochondrial and nuclear genes. Maternally inherited mitochondria (circles) Survivorship 0.55 and nuclear chromosomes inherited from both parents (rods) of populations A and B are colored red and blue, respectively. Crosses produce F1 0.50 hybrids with population A mitochondria. These F1 offspring have slightly higher survival, showing 0.45 “hybrid vigor.” Crosses then produce F2 and F3 hybrids, with recombined nuclear genes. The pa- 0.40 ternal backcross is produced by mating F3 females Parental F1 F3 Paternal Maternal with population B males. These offspring have lines hybrid hybrid backcross backcross low fitness, because most of the nuclear genes come from population B and are mismatched Female to the mitochondrial genes from population A. (population A) In contrast, offspring of the maternal backcross, in which most of the nuclear genes come from the same population as the mitochondria, have Male (population B) normal, high survival. (After.) rearrangements are inversions and reciprocal translocations (see p. 90). Especially in the case of translocations, heterozygotes have reduced fertility compared with Au: Should label “F3 hybrid” be “F2 & F3 hybrid”? This might be a good figure to insert homozygotes fortoeither a balloon caption explainthe original a key point inor the the derived (new) arrangement. For this rea- graph? son, populations with different chromosome arrangements are nearly or entirely monomorphic, and may form narrow hybrid zones where one “chromosome race” meets and interbreeds with another (FIGURE 9.14). The fertility of heterozygotes for chromosome rearrangements may be low either because the rearrangements carry different alleles that create Dobzhansky-Muller incompatibilities, or because of mispairing of chromosomes in meiosis produces gametes that lack certain chro- mosome regions. How fast does reproductive isolation evolve? The time required for reproductive isolation to become strong, after it has started FIGURE 9.14 Two “chromosome races” of the common shrew (Sorex araneus) to evolve, varies greatly. The origin of a new species by polyploidy, which is espe- form a very narrow hybrid zone in Siberia. cially common in plants, requires only one or two generations (see p. 232). If (A) The Novosibirsk and Tomsk races dif- fer by the fusion of some single-armed chromosomes (e.g., o and p in Tomsk) into (A) (B) double-armed chromosomes (e.g., o and Novosibirsk Hybrid Tomsk 1 g in Novosibirsk). In meiosis in hybrids, the Frequency of Novosibirsk type o o multiple rearrangements cause a chain of g 0.8 nine chromosomes to form, and irregular g segregation produces many unbalanced k k 0.6 gametes and low fertility. (B) A transect i i from Novosibirsk to Tomsk shows a cline in h 0.4 the frequency of the Novosibirsk chromo- h some arrangement less than 9 km wide. n n The chromosome configuration of either 0.2 m race cannot increase within populations of m the other race, probably because meiosis p 0 p in F1 hybrids produces gametes that lack 0 5 10 15 20 some chromosomal regions. (A after ; Transect distance (km) B after.) Futuyma Kirkpatrick Evolution, 4e Sinauer Associates Troutt Visual Services  (A) Drosophila (B) Etheostoma Prezygotic isolation 1.0 1.0 Prezygotic isolation appears at Strength of reproductive isolation relatively small genetic distances. 0.8 0.8 Sexual isolation 0.6 0.6 0.4 0.4 0.2 0.2 0 0 0.5 1.0 1.5 2.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Genetic distance Genetic distance Postzygotic isolation 1.0 1.0 Strength of reproductive isolation 0.8 0.8 Hybrid inviability 0.6 0.6 0.4 0.4 0.2 0.2 0 0 –0.2 0.5 1.0 1.5 2.0 0.1 0.2 0.3 0.4 0.5 Genetic distance Genetic distance FIGURE 9.15 Prezygotic isolation evolves faster than postzygotic left part of the two graphs reveals that strong prezygotic isolation isolation in flies and fishes. (A) The strength of prezygotic and evolves shortly after isolation (at small genetic distances), while postzygotic reproductive isolation between pairs of popula- strong postzygotic isolation evolves only later. (B) Similar patterns tions and species of Drosophila increases with the amount of are seen in a genus of freshwater fishes, the darters (Etheos- time since the lineages split. The time is estimated by the genetic toma). Thirteen pairs of allopatric species were tested for both distance between each pair. The strength of prezygotic isolation sexual isolation and the survival of artificially produced hybrids. was measured by observing mating between flies in the labora- For both indices, a value of 0 indicates that the pairs are no more tory. The strength of postzygotic isolation was measured by sur- isolated than conspecific individuals, and a value of 1 indicates vival and fertility of hybrid individuals. Comparison of the upper complete reproductive isolation. (A after ; B after.) 450-plus species of cichlid fishes evolved from 5 ancestral species in Lake Victoria in just 15,000 years (see opening of this chapter), the average time between spe- ciation events was less than 2000 years, which is astonishingly rapid. Molecular clocks (see Chapter 7) can be used to estimate the time back to the most recent common ancestor, giving us the age of their speciation. Based on this approach, sympatric sister species of Drosophila are estimated to have taken about 200,000 years to evolve, while it requires 1.1–2.7 million years for allopatric populations to evolve full reproductive isolation. Some populations of birds that have been diverging for 1.5–3 million years form hybrid zones, showing that it take more time for birds than flies to evolve strong reproductive isolation. In many groups of organisms, prezygotic isolation evolves considerably faster than intrinsic postzygotic isolation (FIGURE 9.15). Consequently, closely related species are often fully interfertile: many species and genera of birds, even after Futuyma Kirkpatrick Evolution, 4e Sinauer Associates more than 5 million years of divergence, can form fully viable, fertile hybrids when Troutt Visual Services crossed. In these cases (which may be the rule in many kinds of organisms), Evolution4e_09.15.ai Date 11-21-2016  Species and Speciation   227 postzygotic isolation probably plays a minor role in speciation. It may, however, affect the further evolution of prezygotic isolation (see p. 230), and it may help keep species separate, because prezygotic barriers such as ecological or sexual isolation may not evolve to completion, or can become weaker if habitats change. For example, increasing turbidity in Lake Victoria interfered with female cichlids’ ability to see differences in male coloration that are the basis of sexual isolation between some closely related species. The result is that species that were previously well isolated are now hybridizing. In contrast, strong postzygotic isolation, such as complete hybrid sterility, is probably irreversible, and can make species permanent. The Causes of Speciation Speciation is the evolution of reproductive isolating barriers. But because these barriers decrease the chance that some individuals mate or that their offspring survive, it might seem paradoxical that they could ever evolve. The solution to this conundrum is that speciation often starts with a geographic barrier (such as a mountain range) that separates two populations of the same spe- cies. Over time, the populations evolve genetic and phenotypic differences, perhaps as they adapt to different ecological conditions. At this stage, there is no reason that genetic differences between the populations, or traits such as mating behavior, should be compatible, because the genes in the two populations are prevented from mixing by the geographic barrier. Sometimes those differences cause prezygotic or postzy- gotic isolation between the populations if they come back into contact (for example, if the mountain range erodes or if colonists disperse across it). If reproductive isolation is sufficiently complete, two species have evolved from one by the process of allopatric speciation. (Remember, we defined reproductive isolation as based on biological dif- ferences that reduce gene exchange, not extrinsic barriers such as mountain ranges.) This scenario illustrates a key point: to initiate speciation, something is needed to restrict free interbreeding between two diverging populations, since interbreed- ing tends to erase their emerging genetic differences. Most often, that restriction results from geographic separation of the populations, although other mechanisms can have this effect. We now turn to the question of what causes the evolution of genetic and phe- notypic differences between geographically separated populations that result in reproductive isolation. That is, what are the causes of speciation? ECOLOGICAL SPECIATION  The two monkeyflower species discussed earlier (see Figure 9.7) provide a vivid example of how reproductive isolation can result when natural selection acts differently on two populations. Based on phy- logenetic reconstruction of ancestral characteristics in the genus, it is likely that the ancestor of these species resembled Mimulus lewisii (see Figure 9.7A): it was bee-pollinated and occupied high elevations. The population that gave rise to M. cardinalis colonized lower elevations, where natural selection favored flower traits that attract hummingbirds: red pigments, abundant nectar, and extension of the petals to form a long, tubular corolla that excludes bees but allows hummingbirds to reach the nectar (see Figure 9.7B). Those changes to the elevational distribu- tion and flowers had the effect of strongly decreasing the exchange of pollen (and genes) with the ancestral population, giving rise to the new species. This scenario is a plausible reconstruction of past events. Biologists have also observed the evolution of reproductive isolation by selection in the laboratory. BOX 9B describes an experiment in which laboratory populations of Drosophila melano- gaster were selected for adaptation to two different environments. In only about 20 generations, the divergently selected subpopulations became substantially repro- ductively isolated. 228   CHAPTER 9 BOX 9B Speciation in the Lab Can different regimes of natural selection cause popula- ments provided conditions for adaptation to the different tions of a species to become different species? Darwin and diets to occur by natural selection. (It was natural selection, many later evolutionary biologists have supposed that this not artificial selection. In artificial selection, the investigator is how speciation usually happens. Indeed, most closely would decide which flies reproduce and which do not. related species have different adaptations to their ecologi- Dodd didn’t do that. Instead, she simply put the flies into a cal circumstances (for example, they often are adapted to stressful environment and let selection take its course.) slightly different habitats or diets), and of course they are After a year, Dodd reared flies from all eight popula- reproductively isolated. But that does not provide evidence tions on standard Drosophila food for one generation (to that the genetic changes underlying their ecological adap- eliminate any maternal effects of starch or maltose). She then tations caused the reproductive isolation. put virgin females and males from a pair of populations to- One way of obtaining relevant evidence is to use experi- gether in an observation chamber and recorded how many mental evolution. In this approach, we expose a laboratory of each of the possible matings occurred. For instance, in population to a simplified version of the conditions we one combination of st and ma populations, two kinds of suspect might occur in nature. The results determine if real “homogamic” matings (female st × male st, female ma × organisms can in principle speciate because of different male ma) and two “heterogamic” matings (female st × male ecological selection pressures. We can also gain other key ma, female ma × male st) might occur. Each of the 16 pos- insights, for example how long the process might take. sible pairs of starch-adapted and maltose-adapted popu- Among many such experiments is one by Diane Dodd lations was tested in this way. In order to be sure than any , who used eight laboratory populations of Drosophila reproductive isolation could be attributed to the divergent pseudoobscura, all of which were founded by flies collect- selection, and not just genetic drift in isolated populations, ed in a single locality in Utah. For 1 year (about 20 genera- Dodd also counted matings between pairs of populations tions), four of the populations were reared on each of two that had been subjected to the same stressful diet. For larval food media, one based on starch (st) and the other every pair of populations, an index of sexual isolation was on maltose (ma). Both media were stressful: Dodd reported calculated that ranged from 0, if the proportion of different- that “it initially took several months for the populations to population matings equaled the proportion of same-pop- become fully established and healthy.” Thus these treat- ulation matings, to 1.0 if no different-population matings The monkeyflowers and the Drosophila experiment illustrate how reproduc- tive isolation can evolve as a side effect of adaptation to different ecological cir- cumstances, a process called ecological speciation [64, 93]. A key point is that the RIBs evolve by pleiotropy (see Chapter 4). There was no direct natural selection for isolation between the populations. Rather, selection acted on other traits that happened to cause isolation. (Recall the distinction between “selection for” and “selection of” features, in Chapter 3.) Although speciation, one of the most impor- tant elements of evolution, is commonly a consequence of adaptive changes in organisms’ characteristics, it is typically not an adaptation itself. SPECIATION BY GENETIC CONFLICT Another powerful cause of the evolu- tion of reproductive isolation is genetic conflict, which occurs when an allele increases its own transmission to the detriment of other alleles at the same or other loci (see Chapter 12). Many mutations have been found that transmit more copies of themselves to the next generation not by increasing survival or repro- duction, but by violating the rules of inheritance. They are transmitted to more than 50 percent of the gametes (a process called segregation distortion). These mutations increase in frequency in a population even though they often reduce  Species and Speciation   229 BOX 9B Speciation in the Lab (continued) occurred. (Incidentally, for these tests Dodd clipped a wing tween mating and rearing environment could have oc- tip on flies from one of the two experimental populations, curred by chance.) But in none of the pairs of populations in order to distinguish them. This procedure did not affect adapted to the same diet was there a statistically significant the results.) excess of same-population matings. Here are the numbers of matings for 1 of the 16 pairs of The sexual isolation index value of 0.46 suggests that in a populations adapted to different diets, and 1 of the 16 pairs mere 20 or so generations, these divergently selected labo- adapted to the same diet: ratory populations had progressed about halfway toward full sexual isolation—in which case speciation would have Different diets (st, ma) Same diet (st) been completed in the laboratory! This is astonishingly fast, Females Females especially in the context of evidence on how long it takes st2 ma2 st1 st2 for speciation to occur in nature (see p. 226). What caused the populations to evolve partial sexual st2 19 7 st1 18 15 isolation? One possible answer is pleiotropy: some of the same genes that enhance adaptation to starch or maltose Males Males might also affect female preference and some feature of ma2 8 22 st2 12 15 males that enables females to distinguish them. Or perhaps the strong selection for alleles that enhance adaptation to Isolation index: 0.46 (P < 0.001) 0.13 the novel diets carried along alleles at closely linked genes that affect male characteristics and female responses to those characteristics. Females of the st and ma populations, adapted to differ- Dodd did not do further research on these possibilities, ent diets, were more likely to accept males adapted to the and in the 1980s it would not have been possible to iden- same diet as themselves. In all 16 combinations of different- tify and obtain the sequences of the relevant genes. That diet populations, there was a tendency for females to show would be a much easier task today. Dodd’s experiment same-diet preference, and this was statistically significant is waiting for someone to repeat it and do the genetic in 11 combinations. (The notation P < 0.001 means that the detective work. probability is less than 1 in 1000 that the correlation be- fertility. Selection therefore favors mutations at other loci that restore full fertility by disabling the segregation distortion caused by the “selfish” mutation. When this conflict between distorter and a restorer has played out in one popu- lation but not another, the populations may be genetically incompatible. This is the basis for strong postzygotic isolation between populations of Drosophila pseu- doobscura in North America and in Bogotá, Colombia: hybrid males are almost completely sterile. Sterility is the result of a mutation at a locus (Overdrive) that reduces male fertility, but that spreads by segregation distortion through the Bogotá population. This population has restorer alleles at other loci that main- tain male fertility, but restoration is inadequate in hybrid males. Genetic con- flict seems to be an important cause of Dobzhansky-Muller incompatibilities in Drosophila and perhaps other groups of organisms. A similar conflict sometimes occurs between nuclear and mitochondrial or chloroplast genes, as in the copepod example described earlier (see Figure 9.13). Earlier we saw that different species of abalones are reproductively isolated because proteins on the outside of their eggs and sperm have diverged to the point where they do not bind to one another. Divergence may have been caused by sexual conflict: changes in the egg surface that slow down sperm entry are  230   CHAPTER 9 3.1 advantageous because fertilization by more than one sperm kills the egg. Any Mean pulse rate of male (pulses/sec) 3.0 such changes in the egg will impose selection for sperm that can beat their 2.9 competitors by penetrating more quickly. 2.8 SPECIATION BY SEXUAL SELECTION  In many groups of rapidly speciat- 2.7 ing animals and plants, species differ more in their secondary sexual traits,1 2.6 such as male coloration or vocalization, than in ecologically important traits (see Figure 9.9). In many cases, one sex (let’s suppose the female) chooses 2.5 mates based on variation in these traits. Females impose strong sexual selec- 2.4 tion, which can drive the rapid evolution of male secondary sexual traits (see 2.3 Chapter 10). Species that differ in sexually selected male features also com- monly differ in female preference, so females recognize and mate preferen- 2.2 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 tially with males of their own species. These patterns suggest that divergent Mean pulse rate preference of female sexual selection can cause rapid evolution of prezygotic isolation between (pulses/sec) populations. Certain groups of animals, such as cichlid fishes and hum- mingbirds, have indeed speciated rapidly and show strong sexual selection. FIGURE 9.16 Evolution of sexual isolation by Recent phylogenetic analyses of birds suggest that male coloration patterns sexual selection. The pulse rate of the mating call of male crickets (Laupala cerasina) and associated both with sexual selection and species recognition evolve fastest the pulse rate preferred by females both vary in lineages with high speciation rates. among local populations. These differences Studies of closely related populations and species provide more direct evi- are genetically based. The confidence intervals dence that sexual selection may cause speciation. For example, male calls and around each point show that the preference female preferences covary among populations of a Hawaiian cricket (Laupala ranges of females of the most widely differ- cerasina), to the point that females hardly respond to the calls of the most dif- ent populations would not include the most ferent population (FIGURE 9.16). Sexual isolation appears to be the sole divergent males. (After.) basis of reproductive isolation between some ecologically indistinguishable species of freshwater fishes called darters (see Figure 9.15B). Why then does sexual selection vary among populations? In Chapter 10 we will con- sider some of the factors at work. These include direct benefits to mate prefer- ences, selection acting on pleiotropic effects of preference genes, preferences for mates with “good genes,” and ecological factors that make different courtship signals more effective in different environments. REINFORCEMENT OF REPRODUCTIVE ISOLATION  So far, we have discussed how speciation can result as a side effect of divergent selection. In some cases, natural selection can also directly favor the evolution of prezygotic isolation. Consider two populations that have already evolved some degree of isolation so that hybrids have lower survival or fertility. A female that chooses a male from her own population will leave more descendants than one that makes the mis- take of mating with a male from the other population. This creates a selective advantage to an allele for a mating preference that increases the chance of mating within rather than between populations. A “discrimination” allele will be trans- mitted to more progeny, on average, than a “random-mating” allele. The evolution of stronger prezygotic isolation because of selection against low- fitness hybrids is called reinforcement. Not all types of isolating mechanisms can evolve this way. Alleles that strengthen prezygotic isolation gain an advantage because individuals with them have higher fitness than do those that hybridize. But stronger postzygotic isolation usually cannot evolve by natural selection. An allele that lowers hybrid fitness cannot increase in frequency, for that would be the antithesis of natural selection. (Exceptions are in organisms such as plants and mammals, in which embryos compete for the mother’s nutrients. It can be advan- tageous for a mother to abort hybrid embryos and allocate resources to nonhybrid 1 Secondary sexual traits are those that differ between the sexes, other than the gonads and re- productive structures. Futuyma Kirkpatrick Evolution, 4e Sinauer Associates Troutt Visual Services  Species and Speciation   231 (A) (C) 60 Texas 50 100 km P. cuspidata 40 P. drummondii Fruit set Austin 30 20 P. drummondii 10 0 Light blue Light red Dark blue Dark red (B) Allopatry Sympatry 60 50 Relative hybridization 40 Light red 30 (iihh) 20 Light blue Dark red (IIhh) 10 (iiHH) 0