Natural Selection may take a variety of forms and act on any behavioral, morphological, developmental or physiological traits of an organism. However, certain types of selection are unique in their features, and they are often treated as special categories of selection. One of these "special" categories is that of sexual selection.
Darwin was the first to realize the existence and importance of sexual selection, which he defined as "the advantage which certain individuals have over others of the same sex and species solely in respect of reproduction". The modern concept of sexual selection is usually restricted to characteristics which affect mating success, not reproduction in general, but most organisms must mate in order to reproduce!
Darwin (1871) developed the notion of sexual selection because he realized that not all of the differences between males and females were due to their specialized roles in sexual reproduction. Features found in males of the species such as horns or antlers (or disproportionately larger ones) in ungulates (e. g., deer, gazelles), significantly larger body size in many mammals, or the bright plumage common in birds have nothing to do with sperm production or the physical aspects of mating. In fact, Darwin realized, some characteristics, like bright plumage, might be detrimental to an individual because they hampered the organism or made it more susceptible to predation. However, such characteristics would be favored if the enhancement of mating success outweighed (or at least equaled) their detrimental effects.
Sexual Selection may take two forms: (1) Intrasexual Selection involves characteristics which affect the outcome of competition among members of one sex for access to members of the other sex. For example, intrasexual selection would operate on physical and behavioral features which helped to determine the outcome of aggressive encounters among males over territories, if possession of or the quality of a territory affected subsequent mating success. (2) Intersexual Selection would influence the evolution of secondary sexual characteristics which determine the relative "attractiveness" of members of one sex to the other sex. Such items as courtship displays and male plumage in birds (e. g., the male peacock) are obvious examples, but the whole phenomenon of mate choice may often be subtle (see Bateson, 1983). Often, it is difficult to separate the two forms of sexual selection. The earlier case in which males competed among themselves for high quality territories, which would then determine the number of females who mated with each male, is clearly a case where there are components of both intrasexual and intersexual selection.
Sexual Selection is also noteworthy because it has the potential to become a "runaway" process in which there is no obvious optimum set of characters. In other categories of selection there is at least the potential for some defined equilibrium point. However, in sexual selection there is no optimum as long as the potential exists for further improvements in mating success. Take the hypothetical case where a female bird chooses her mate based on some plumage characteristic like tail feather length. The subsequent male offspring will inherit the tail feather length, but the female offspring will inherit the mate preference behavior so that the cycle of selection will continue in the next generation. The process of sexual selection has the potential to become a case of positive feedback in which the characteristic is continually enhanced in an (almost) endless cycle.
R. A. Fisher was the first to define the possibility that sexual selection could become a "runaway" process. However, he did not examine this process in explicit, quantitative genetic detail. Recent models by Russell Lande have shown that the runaway process is not inevitable (see review by Arnold, 1987). According to Lande's work, the runaway process will be most likely to occur if: (1) there is a high genetic correlation between male and female traits (i. e., a high genetic covariance); (2) there is weak natural selection on the male characters; (3) there is a strong mating preference by the females. The runaway process is in fact an example of an unstable equilibrium, but the stable case, which occurs when there is weak intersexual correlation, strong natural selection, and a "weak" mating preference, has some interesting properties. In some models, the stable equilibrium is not a single point on the adaptive topography, but instead it is a line of equilibrium values. In other words, it represents a neutral equilibrium, with the population's final location being affected by other factors - such as genetic drift (see figure). However, regardless of the nature of the equilibrium, most models show that the population may reach an equilibrium below the peak of the adaptive topography because the balance between sexual selection and other selective processes can reduce the average fitness of the population.
Regardless of the form of equilibrium, the process of sexual selection is limited. Sexual selection will cease to affect a character when one or more of the following occur: (1) There is no longer any genetic variation for the characteristic; all individuals are both genotypically and phenotypically identical. (2) There is still genetic variability, but it results in phenotypic differences which cannot be perceived by other individuals, and some other characteristic becomes the basis of sexual selection. (3) The effects of sexual selection are counterbalanced by some other form of selection. For example, there may be plenty of potential for brighter male plumage, but any increase in male "showiness" would raise the risk of predation to the point where a male would be unlikely to survive reaching sexual maturity.
Sexual selection usually operates on the sex with the greater degree of variability in mating success. Most of the documented cases of sexual selection focus on males because, presumably, all females will be mated, but males will vary in their mating success. This is especially true in polygynous species where a male may control mating with a group of females (sometimes such males are described as having a harem), and so some males make a disproportionate contribution to the next generation, while some males do not mate at all. However, there are cases in which sexual selection acts on females because the males make a reproductive contribution which varies so that some males are more "desirable" mates than others (e. g., Petrie 1983, 1986), and there is female competition for these males.
The discussion so far has avoided the issue of why sexual selection for a particular character should exist at all. While intrasexual selection may be easily understood and observed, the phenomenon of intersexual selection is more subtle because it is not clear why, or how, a particular feature became the basis by which mates are chosen. That there is a genetic basis to mate preferences has been demonstrated in a number of cases (Majerus, 1986). Less certain, however, is if there is any genetic correlation between the sexually selected trait and any component of mate fitness, or fitness of the subsequent offspring. In fact, some have argued that the sexually selected trait has its own intrinsic value because of the enhanced mating success of the offspring, but the so-called "sexy son" hypothesis does not hold up under quantitative genetic scrutiny (Arnold, 1987).
The fact remains that we have few good examples of the "good genes" hypothesis, which holds that there is a correlation between the intersexually selected trait and the fitness of subsequent progeny. One hypothesis, by Hamilton and Zuk (1982), proposes that sexually selected characters are correlated with resistance to parasites. This hypothesis is based on the fact that parasitism is a prevalent, powerful selective agent, and if resistance to parasites is a heritable characteristic, correlated with a qualitative secondary sexual characteristic (e. g., plumage color), then this could be a case that fulfills the "good genes" scenario. Although there are specific requirements associated with the Hamilton and Zuk hypothesis, there is some empirical evidence to support the role of parasitism in sexual selection. (Read, 1988).
Sexual selection may also play a role in the process of speciation because mate preferences are obvious foundations for premating reproductive isolating mechanisms. Russell Lande (1981, 1982) has investigated this possibility. Given the possibility that sexual selection can result in linear, rather than point equilibria, it is easy to imagine two populations which have the same intersexual system, but they attain different equilibria as a result of genetic drift and/or slight differences in initial conditions. Presumably, individuals moving from one population to the other would experience reduced fitness due to decreased mating success. Lande (1982) has also shown that sexual selection may exaggerate clinal variation, which would further increase the likelihood of speciation.
The most intriguing hypothesis to develop from recent discussions of the importance of sexual selection is William Eberhard's proposal regarding the evolution of animal genitalia (Eberhard, 1985, 1990). In many species of animals, especially insects, external genital structures are both elaborate and species specific. The classical explanation is that these structures act as pre-mating reproductive isolating mechanisms - the "lock and key" hypothesis - requiring proper matching for the transfer of gametes. However, precise physical matching of genitalia is not necessary for successful gamete transfer (Eberhard, 1990). Instead, he argues that the external genitalia are part of an elaborate premating system that is necessary to guarantee fertilization. In other words, the genital structures have evolved through a subtle form of sexual selection. This means that the potential range of sexual selection should be extended to many, if not all, prezygotic phenomena.