This Article Was Originally Published in the Encyclopedia of Animal

This article was originally published in the Encyclopedia of Animal Behavior published by Elsevier, and the attached copy is provided by Elsevier for the author's benefit and for the benefit of the author's institution, for non- commercial research and educational use including without limitation use in instruction at your institution, sending it to specific colleagues who you know, and providing a copy to your institution’s administrator.

All other uses, reproduction and distribution, including without limitation commercial reprints, selling or licensing copies or access, or posting on open internet sites, your personal or institution’s website or repository, are prohibited. For exceptions, permission may be sought for such use through Elsevier's permissions site at:

http://www.elsevier.com/locate/permissionusematerial

Grether G.F. (2010) Sexual Selection and Speciation. In: Breed M.D. and Moore J., (eds.) Encyclopedia of Animal Behavior, volume 3, pp. 177-183 Oxford: Academic Press.

Author's personal copy

Sexual Selection and Speciation

G. F. Grether, University of California, Los Angeles, CA, USA

Historical Background It would be hard to deny that sexual selection is at least

partly responsible for the extraordinarily high rates of Sexual selection was Charles Darwin’s solution to the exis- speciation in African lake cichlid fishes, in which some tence of secondary sexual characters, such as the ornate closely related species differ only in male coloration and plumage of male birds of paradise, the exaggerated weap- female color preferences, or in Hawaiian crickets, in onry of male horned beetles, and the elaborate antennae of which morphologically indistinguishable sister species male gypsy moths. These traits were puzzling, precisely differ in male song and female song preferences. Exactly because they seemed costly for survival. Darwin’s explana- what role sexual selection has played in the radiation of tion was that such traits evolve in spite of their survival these and other taxonomic groups is an open question. costs because they increase male mating success. He pro- This is a very active area of research. posed that male ornamentation evolved in response to female mate choice, that male weaponry evolved because of contest competition among males for access to females, Processes Linking Sexual Selection and that enlarged male sensory organs evolved because of and Speciation scramble competition for females.

Sexual selection by female choice, arguably Darwin’s What distinguishes sexual selection from most other evo- most controversial idea, was criticized by several promi- lutionary processes is its potential to cause rapid prezy- nent evolutionists, including Alfred Russell Wallace. gotic or behavioral isolation. Although sexual selection

Darwin himself seemed unable to explain why females can also increase postzygotic isolation by reducing the should prefer ornamented males. Ronald Fisher outlined mating success of hybrids, this requires some degree of the basic explanations that are widely accepted today, but prezygotic isolation. I, therefore, restrict my attention it took another 50 years for these ideas to be formally here to processes leading to prezygotic isolation (with or modeled and tested empirically. In the meantime, sexual without postzygotic isolation). selection was supplanted by the notion that secondary Whether a given mode of speciation is viewed as driven sexual characters are reproductive isolating mechanisms, by sexual selection depends on how narrowly the term that is, traits that prevent interbreeding between closely sexual selection is interpreted. Strictly speaking, sexual related species. selection refers to covariation between traits and mating These two ways of thinking about secondary sexual success, where mating success includes the quantity and/ characters were formally united in the 1980s. Mathematical or quality of mating partners. However, mate preferences modelsby Russell Lande and other evolutionary theorists and secondary sexual characters may also evolve in showed how sexual selection, in combination with genetic response to other forms of natural selection, with poten- drift or ecological gradients, could cause populations to tially similar consequences for reproductive isolation and diverge in mate preferences and secondary sexual charac- speciation. If mate preferences and secondary sexual char- ters tothe point of reproductive isolation. Research on acters are genetically correlated (as assumed under some sexual selection exploded in the 1980s, but with few excep- sexual selection models), then selection on one would tions, mate recognition between species and mate choice yield a correlated response in the other, and thus, geo- within species continued to be treated as separate phenom- graphic variation in the strength of survival selection could ena. After the first wave of studies established the ubiquity result in a pattern of correlated variation in secondary of female choice, finally vindicating Darwin, the primary sexual traits and mate preferences. While theoreticians focus of empirical research began to test the assumptions have tended to focus on speciation via sexual selection in behind several alternative models of mate preference evo- the strict sense, most empirical studies (especially compar- lution. Empirical research on the role of sexual selection in ative studies) are unable to identify the specific mechan- speciation was uncommon until the last few years of the isms of selection. More to the point, there are many twentieth century. possible causal links between sexual selection and specia- Since speciation was studied for decades without any tion, and the links that are most often explored by theore- explicit consideration of sexual selection, it is worth ask- ticians may not be the most prevalent in nature. ing whether taking sexual selection into account is really At least eight processes could contribute to prezygotic necessary. Ultimately, this is an empirical question. reproductive isolation by causing population divergence in

177 Encyclopedia of Animal Behavior (2010), vol. 3, pp. 177-183

Author's personal copy

178 Sexual Selection and Speciation secondary sexual characters and mate choice (see Table 1). other population mate indiscriminately, are likely to cause Of these, only reinforcement represents direct selection for asymmetrical gene flow but not reproductive isolation. Like- reproductive isolation; under the other processes, repro- wise, shifts in secondary sexual characters that are not ductive isolation emerges as a byproduct. Reinforcement accompanied by corresponding shifts in mate preferences, can be viewed as a form of good genes sexual selection, are unlikely to cause reproductive isolation (e.g., males from where members of the other population represent low the population with the most elaborate version of the char- quality mates (i.e., hybrid offspring have low fitness), but acter may simply be preferred by females from both popula- it is usually treated separately from sexual selection. In tions). In general, shifts in sexual traits or mate preferences reinforcement models, sexual selection can be side-stepped alone are unlikely to cause speciation. by assuming the existence of a single assortative mating (‘like mates with like’) locus, although it is probably more realistic to model mate preference and secondary sexual Mathematical Models of Speciation by character loci separately. Aside from reinforcement, Fish- Sexual Selection erian selection and sexual conflict have received the most theoretical attention. The remaining processes listed in Most models of speciation by sexual selection are based

Table 1 represent relatively unstudied links between sex- on Lande’s formalization of Fisher’s insight that mate ual selection and speciation that probably deserve greater preferences can coevolve merely because both sexes attention. None of these processes is mutually exclusive; carry genes that influence the expression of mate prefer- multiple processes operating together may be more likely ences and secondary sexual characters (process 3 in to cause speciation than any single process. Table 1). Females with a strong preference for, say, long- Not all instances of divergence in secondary sexual char- tailed males, tend to produce offspring with genes for both acters and mate preferences are equally likely to cause long tails and the long-tail preference. This can result in a speciation. Asymmetrical mate preferences, in which positive feedback loop in which mate preferences and females of one population mate assortatively (i.e., prefer secondary sexual characters coevolve in unpredictable malesfrom their natal population), while females from the ways, potentially leading to reproductive isolation be-

tween allopatric populations. Populations can diverge by Table 1 Processes that could contribute to reproductive drift alone, but the prospects for speciation are enhanced isolation by causing populations to diverge in secondary sexual when populations occupy different positions along an characters and mate choice environmental gradient that influences the strength of Process Definition in this context survival selection on the male character. Under some circumstances, Fisherian selection could 1. Genetic drift Shifts in secondary sexual characters or result in sympatric speciation. This has been the subject of mate preferences caused by genetic several modeling efforts inspired by the African lake bottlenecks or founder events cichlids. A frequent conclusion is that sympatric specia- 2. Reinforcement Evolution of increased prezygotic tion is more likely when sexual selection is coupled with reproductive isolation between sympatric populations in response to selection ecological (niche) divergence. Models based only on dis- against hybrids ruptive sexual selection yield sympatric speciation under 3. Fisherian Co-evolution of male secondary sexual more restrictive conditions than models that also allow selection characters and female preferences ecological character displacement to occur. arising from the genetic correlation between the sexes Models by Sergey Gavrilets and colleagues suggest 4. Sexual conflict Co-evolution of male secondary sexual that the sexual conflict mode of sexual selection may be characters and female resistance to particularly likely to result in speciation (process 4). In mating these models, mating is costly for females (above some 5. Intrasexual Divergence in secondary sexual characters optimal mating rate) and female preferences arise from competition caused by agonistic interactions (e.g., resistance to mating. Males evolve adaptations to over- territoriality) 6. Local Selection arising from shifts in the local come female resistance and females evolve counteradap- adaptation optima of secondary sexual characters or tations. Allopatric populations can rapidly diverge to the mate preferences (e.g., sensory drive, point where males are unable to mate with females from predation) another population, resulting in reproductive isolation. 7. Pleiotropy or Shifts in secondary sexual characters or linkage mate preferences caused by selection on Speciation is not an inevitable outcome, however; males genetically correlated traits from one population could be superior at mating with 8. Phenotypic Within-generation shifts in the development females from both populations (asymmetric preference). plasticity or expression of environmentally Sexual conflict can also cause sympatric speciation if sensitive secondary sexual characters or females evolve two alternative strategies for resisting mate preferences mating and males evolve adaptations for specializing on

Encyclopedia of Animal Behavior (2010), vol. 3, pp. 177-183

Author's personal copy

Sexual Selection and Speciation 179 one type of female or the other, but the likelihood of this received some support. Pitch and temporal patterning of outcome is unclear. songs correlated with body mass and bill size, respectively,

From an empirical standpoint, the sobering message as predicted from biomechanical constraints on song pro- from theoretical work is that many different outcomes duction (i.e., pleiotropy). Pitch also correlated with acous- are possible, even within the relatively small subset of tic transmission properties of the forest strata in which parameter space that has been explored so far. Still, it antbirds typically sing, as predicted by the acoustic adap- may be possible to parameterize models for particular tation hypothesis. Finally, closely related sympatric spe- systems to evaluate whether speciation is a likely outcome. cies differed more in song than closely related allopatric

Future theoretical work could also clarify which assump- species, providing evidence for reinforcement or postspe- tions of the models most strongly affect the prospects for ciation reproductive character displacement. Whether speciation. divergence in song prevents interbreeding between ant- bird species is unknown, however; this would be a useful

direction for future research. Empirical Evidence

Overview Case Studies

A wide range of data have been put forth as evidence that The East-African lake cichlids have become an iconic sexual selection plays a role in speciation. Here, I review example of speciation by sexual selection. In Lake Victo- three categories of evidence: taxonomically broad com- ria, for example, over 500 species appear to have evolved parative studies (family level or above), case studies of from a few ancestral species in the past 100 000 years. smaller taxonomic scale, and experimental evolution Many of the species differ primarily in male coloration and are genetically isolated from each other only by studies. female preferences. Hybrids are viable, fertile, and inter- Broad Comparative Studies mediate in coloration. Some species contain multiple male color morphs and morph-specific female prefer-

Several published studies correlating species richness ences, and may be in the process speciating yet again. with putative indices of sexual selection (e.g., sexual Although several models of speciation have been loosely dimorphism, degree of polygyny, size of testes) have based on this system, it would be wrong to infer that we found positive correlations, after controlling for phylog- understand how speciation occurs in the African lake eny. Similar studies, however, have found no correlations cichlids themselves. How color polymorphisms arise and or even negative correlations. Positive correlations sug- how they are maintained long enough for reproductive gest that sexual selection increases speciation rates or isolation to evolve, is largely an unsolved mystery. One decreases extinction rates. There are several reasons to solution proposed by Ole Seehausen and Dolph Schluter suggest that sexual selection might increase extinction is that intrasexual (male–male) competition over breeding rates but few, if any, reasons to expect the reverse. territories favors rare color morphs (process 5), setting the

Hence, these results have been taken as evidence that stage for the evolution of morph-specific mate prefer- sexual selection increases speciation rates. One important ences and reproductive isolation. Several indirect lines caveat is that indices of sexual selection may correlate of evidence support this hypothesis, but direct tests for with other factors that affect species richness. Taxonomy color-based aggression biases have yielded conflicting is often based on male secondary sexual characters and results. An alternative hypothesis is that the color diver- allopatric populations are more likely to be classified as gence is a product of small-scale differences between separate species if they differ in such characters. Thus, species in breeding habitat, coupled with sensory drive speciesrichness might be systematically overestimated in (process 6). Water color is red-shifted (i.e., shifted towards clades with elaborate genitalia, complex song, or bright long wavelengths) at greater depths, and thus, different coloration. Consequently, using sexual dimorphism in colors are conspicuous at different depths. In one sympat- such traits as an index of sexual selection is problematic, ric species pair, the species in which males are typically and to the extent that other indices of sexual selection yellow and red (Pundamilia nyererei ) breeds at greater correlate with sexual dimorphism, they may suffer from depths than the species in which males are typically the same problem. Another reason to be leery is that blue (P. pundamilia)(Figure 1), and females use male similar studies on the same taxonomic groups (e.g., color to mate assortatively by species. In optomotor birds) have yielded contradictory results. tests, female P. nyererei are more sensitive to red light In a phylogenetic analysis relating song to morphology and female P. pundamilia are more sensitive to blue light. across 163 species of antbirds (Thamnophilidae), Nathalie These species differences in wavelength sensitivity can be Seddon tested predictions based on pleiotropy, local adap- largely explained by sequence divergence in genes coding tation, and species recognition. All three hypotheses for visual pigments (opsins). Seehausen and colleagues

Encyclopedia of Animal Behavior (2010), vol. 3, pp. 177-183

Author's personal copy

180 Sexual Selection and Speciation

1600

1200

800

Intensity

400

0 400 450 500 550 600 650 700 750

(b) Wavelength (nm)

(a)

Figure 1 (a) Variation in male nuptial coloration in sympatric Pundamilia spp. cichlids of Lake Victoria, from the blue form of typical Pundamilia pundamilia (top) to the red form of typical P. nyererei (bottom). Females of both species have cryptic yellowish coloration. The red species P. nyererei is found at greater depths than the blue species P. pundamilia. (b) An illustration of how the spectrum of ambient light changes with depth in Lake Victoria, from the surface (blue) through three successive depths: 0.5 m (green), 1.5 m

(orange), and 2.5 m (red). Modified and reprinted by permission from Seehausen O, Terai Y, Magalhaes IS, et al. (2008) Speciation through sensory drive in cichlid fish. Nature 455: 620–626, with permission from Nature Publishing Group. Copyright Macmillan

Publishers Ltd.

have proposed a complex verbal model of the speciation Three-spine sticklebacks provide an example in which process that includes sensory drive, Fisherian selection, sexual selection and ecological character displacement gene flow, and reinforcement. both appear to have played integral roles in speciation. Hawaiian crickets in the genus Laupala have the high- In the lakes of British Columbia, sticklebacks occur in two est speciation rate on record for anthropods. Closely ecologically and morphologically distinct species pairs: a related Laupala species are morphologically and ecologi- larger benthic ecotype that forages in the littoral zone, cally indistinguishable and can produce viable hybrids. and a smaller limnetic ecotype that forages in open water. The only conspicuous difference between sympatric spe- Limnetic and benthic ecotypes within a lake are more cies is that they differ in the pulse rate of male courtship closely related to each other genetically than they are song. Tamra Mendelson and Kerry Shaw found that to fish of the same ecotype in different lakes, proba- female Laupala can discriminate between conspecific and bly because each lake was colonized independently by heterospecific song from a distance and are more likely to the marine ancestor (Gasterosteus aculeatus). Nevertheless, approach conspecific males. This suggested that corre- benthics and limnetics within a lake are reproductively lated divergence in song and song preferences drove spe- isolated, while fish of the same ecotype from different ciation, but further research by the same researchers lakes are not. This suggests that the same prezygotic revealed a more complex story. In a laboratory study of isolating barriers arose independently in different lakes. two allopatric species, conspecific courtship sequences Indeed, Janette Boughman and colleagues have shown usually went to completion, while heterospecific court- that male coloration and female sensitivity to red light ship rarely proceeded to the stage where males provide differ between ecotypes in the same direction in three spermatophores. However, when females of the same spe- different lakes. Compared to benthics, limnetic males cies were paired with F2 hybrid males, which vary widely have more red and less black coloration, and limnetic in song pulse rate, the song pulse rate of the males did not females are more sensitive to red light. In each lake, predict whether courtship proceeded to completion. reproductive isolation between ecotypes appears to be A possible explanation is that chemical or tactile cues maintained by female choice based on male size and are exchanged between the sexes during courtship and color. The consistent direction of the differences between that divergence in such cues, not courtship song, is ecotypes in male color and female sensitivity to red light responsible for the breakdown in heterospecific courtship. suggests that they are caused by habitat differences (albeit In support of this explanation, the researchers found in the opposite direction as seen in African lake cichlids). evidence for rapid divergence between Laupala species Water color is red-shifted at greater depths in these lakes, in cuticular hydrocarbon (CHC) profiles. Whether spe- and male benthics raised in red-shifted water in the labo- cies differences in CHCs contribute to reproductive iso- ratory develop less red and more black coloration than lation remains to be determined. those raised in clear water. It would be informative to

Encyclopedia of Animal Behavior (2010), vol. 3, pp. 177-183

Author's personal copy

Sexual Selection and Speciation 181 know whether this response to ambient light is present cause rapid speciation (process 8). Indigobird nestlings, in the marine stickleback (i.e., the presumed ancestor). which are invariably raised by foster parents, imprint on

If so, and if the sensitivity of females to red light is the songs of their host species. Males later attract females similarly affected, then reproductive isolation between reared by the same host species by mimicking host songs the ecotypes might have arisen as a byproduct of plastic (Figure 2), and females preferentially lay eggs in the nests responses to the environment (process 8). Alternatively, of their host species. Normally, this process of sexual or in addition, the ecotype differences in color and red imprinting maintains host-specificity between genera- sensitivity might have evolved in response to selection, tions. But when females lay eggs in the nest of species favoring increases in the visibility of males against the other than their natal host, sexual imprinting may result background water color (process 6). Reinforcement in the sudden formation of new host races, or hybridiza- seems unlikely in this case because ecotypes differ only tion between existing host races, depending on whether in the strength and not in the direction of the female color the novel host already has its own host race of indigobirds. preference. Molecular genetic data support this model of sympatric Research on brood-parasitic indigobirds (Vidua spp.) by speciation with occasional hybridization. As a possible Michael Sorenson, Robert Payne, and colleagues illus- example of speciation in action, one of the ten recognized trates how within-generation shifts in the development parasitic indigobird species occurs in two morphologi- of secondary sexual traits and mate preferences could cally indistinguishable host races. Males sing host-specific

Mimicry of L. senegala Mimicry of L. striata

(a) (g)

(b) (h)

(d) (j)

(e) (k) 8

6 4 2

Frequency (kHz) (f) (l)

0 1.0 2.0

Time (s) Figure 2 Sonograms showing host mimicry by male indigobirds (Vidua chalybeata). Males represented in the left column were reared by the normal firefinch host, while those represented in the right column were reared by Bengalese finches. Reprinted with permission from Payne RB, Payne LL, Woods JL, and Sorenson MD (2000) Imprinting and the origin of parasite–host species associations in brood-parasitic indigobirds, Vidua chalybeata. Animal Behaviour 59: 69–81, with permission from Elsevier. Copyright Elsevier.

Encyclopedia of Animal Behavior (2010), vol. 3, pp. 177-183

Author's personal copy

182 Sexual Selection and Speciation songs and respond more aggressively to songs of their overcome female resistance and female counteradaptations own host race. Whether the host races are reproductively may evolve in parallel, from a common starting point, as isolated remains to be determined. opposed to diverging between selection lines. If so, this Other study systems that have provided valuable would imply that sexual conflict alone is unlikely to cause insights into the role of sexual selection in speciation speciation, at least not in Drosophila. But perhaps sexual include Galapagos finches, Hawaiian Drosophila, Pacific conflict would cause divergence between populations that salmon (Oncorhynchus), greenish warblers (Phylloscopus), inhabit environments that differ in ways that affect the cost andwater striders (Gerridae), among others. See the review of female resistance or the efficacy of male mating tactics, article by Ritchie for discussions on most of these examples. etc. One promising direction for future research would be to incorporate such environmental gradients into experi- Experimental Evolution Studies mental evolution studies.

In this context, experimental evolution refers to the approach of establishing replicate laboratory populations Conclusions and Future Directions or lines, maintaining them under specific treatments that are thought to represent different intensities of sexual Research on speciation, at least in animals, needs to take selection (e.g., monogamy vs. promiscuity), and testing sexual selection into account. On the other hand, the quest for reproductive isolation between lines after multiple for examples of speciation, driven purely by sexual selec- generations. This method has the potential to demon- tion, is misguided. Theoretical work shows that speciation strate the efficacy of particular scenarios for generating is more likely when divergent sexual selection is coupled reproductive isolation. So far, this approach has only been with ecological divergence, and empirical studies suggest used to test the sexual conflict model. An experimental that the situation in nature is even more complex. While it evolution study with dungflies showed incipient repro- is generally productive in science to pit alternative hypoth- ductive isolation (assortative mating) after 35 generations eses against each other, in this case, the hypothesized pro- under conditions promoting sexual conflict (Figure 3), cesses may be integral parts of a more complex process. but similar experiments on Drosophila melanogaster and A goal for future modeling should be to evaluate which D. pseudoobscura failed to yield any signs of reproductive combinations of processes most readily yield speciation. isolation after 40–50 generations. It is often said that we need only a couple of good One possible explanation for the negative results in examples of a given evolutionary process, but in the case Drosophila is that replicate populations exposed to the of sexual selection and speciation, no well-studied sys- same selection regimes do not evolve in arbitrarily different tems are directly comparable. We certainly have not directions, as modelers have assumed; that is, adaptations to reached the stamp-collecting stage of research on this

topic. There are many well-documented examples of allopatric populations that have diverged in mate prefer- 0.7 ences or secondary sexual characters, but most such stud-

0.6 ies stop short of determining whether the populations are Monogamy reproductively isolated and, if so, whether the observed 0.5 phenotypic differences are responsible. The Hawaiian crickets illustrate the problem with assuming that repro- 0.4 ductive isolation is caused by an observed difference in 0.3 Low density traits and preferences. Future case studies should go beyond identifying plausible mechanisms of prezygotic 0.2 isolation to showing whether these mechanisms actually High density operate in nature. Another promising direction for future

Proportion of pairs copulating 0.1 research will be to use the experimental evolution

0 Between populations Within populations approach to identify the conditions under which particular modes of sexual selection generate reproductive isolation. Figure 3 Results of experimental evolution study on dungflies. Behavioral ecologists have recently begun to pay more The graph shows incipient reproductive isolation between populations (lines) of flies maintained under conditions attention to species in which females develop secondary conducive to sexual conflict (low-density and high-density sexual characters. It is not clear yet whether female- treatments) but not between populations maintained under specific secondary characters are products of sexual selec- Æ monogamy, after 35 generations. Error bars indicate 1 standard tion or some other form of social selection (e.g., resource error. Reprinted by permission from Martin OY and Hosken DJ (2003) The evolution of reproductive isolation through sexual competition). The role of female secondary sexual char- conflict. Nature 423: 979–982, with permission from Nature acters and male mate choice in speciation is a wide open Publishing Group. Copyright Macmillan Publishers Ltd. question.

Encyclopedia of Animal Behavior (2010), vol. 3, pp. 177-183

Author's personal copy

Sexual Selection and Speciation 183

See also: Isolating Mechanisms and Speciation; Mate Gavrilets S and Hayashi TI (2005) Speciation and sexual conflict. Choice in Males and Females; Mating Signals; Parasites Evolutionary Ecology 19: 167–198. Lande R (1981) Models of speciation by sexual selection on polygenic and Sexual Selection; Social Selection, Sexual Selection, traits. Proceedings of the National Academy of Sciences of the and Sexual Conflict. United States of America 78: 3721–3725. Lewandowski E and Boughman JW (2008) Effects of genetics and light environment on colour expression in threespine sticklebacks. Biological Journal of the Linnean Society 94: 663–673. Further Reading Mendelson TC and Shaw KL (2005) Sexual behaviour: Rapid speciation in an arthropod. Nature 433: 375–376. Mullen SP, Mendelson TC, Schal C, and Shaw KL (2007) Rapid Andersson M (1994) Sexual Selection. Princeton, NJ: Princeton evolution of cuticular hydrocarbons in a species radiation of University Press. acoustically diverse Hawaiian crickets (Gryllidae: Trigonidiinae: Bacigalupe LD, Crudgington HS, Hunter F, Moore AJ, and Snook RR Laupala). Evolution 61: 223–231. (2007) Sexual conflict does not drive reproductive isolation in Ritchie MG (2007) Sexual selection and speciation. Annual Review of experimental populations of Drosophila pseudoobscura. Journal of Ecology Evolution and Systematics 38: 79–102. Evolutionary Biology 20: 1763–1771. Seddon N (2005) Ecological adaptation and species recognition drives Balakrishnan CN and Sorenson MD (2006) Song discrimination vocal evolution in neotropical suboscine birds. Evolution 59:

suggests premating isolation among sympatric indigobird species 200–215.

and host races. Behavioral Ecology 17: 473–478. Seehausen O and Schluter D (2004) Male–male competition and

Boughman JW, Rundle HD, and Schlutes D (2005) Parallel evolution of nuptial-colour displacement as a diversifying force in Lake Victoria

sexual isolation in sticklebacks. Evolution 59: 361–373. cichlid fishes. Proceedings of the Royal Society of London Series B: Coyne JA and Orr HA (2004) Speciation. Sunderland: Sinauer Biological Sciences 271: 1345–1353. Associates. Seehausen O, Terai Y, Magalhaes IS, et al. (2008) Speciation through Darwin C (1871) The Descent of Man, and Selection in Relation to Sex. sensory drive in cichlid fish. Nature 455: 620–626. London: John Murray. Sorenson MD, Sefc KM, and Payne RB (2003) Speciation by host switch Endler JA (1992) Signals, signal conditions, and the direction of in brood parasitic indigobirds. Nature 424: 928–931. evolution. American Naturalist 139: 1–27.

Encyclopedia of Animal Behavior (2010), vol. 3, pp. 177-183