Chapter One
The Opportunity for Selection
"When the males and females of any animal have the same general habits of life, but differ in structure, colour, or ornament, such differences have been mainly caused by sexual selection." -(Darwin 1859, p. 89)
Sexual Selection and the Sex Difference in Variance of Reproductive Success
Darwin recognized two patterns in nature and used them to frame the central questions of sexual selection (Darwin 1859):
1. Why do males and females of the same species differ from one another, with males exhibiting morphological and behavioral phenotypes more exaggerated than those of females?
2. Why do the males of closely related species exhibit much greater differences in morphology and behavior than the females of closely related species?
The first pattern is a microevolutionary one, seen commonly within species of almost all taxa with separate sexes, including plants. It indicates that some kind of selection is working to differentiate the sexes and it is affecting males to a much greater degree than females. The second pattern is a macroevolutionary one, observed across species within genera or families of almost all taxa (e.g., many avian taxa). These large differences in male phenotype among closely related taxa are the signature of a very strong and rapid evolutionary force. Darwin noted that, in many species, the phenotypic differences between the sexes are associated neither with essential reproductive physiology nor with development of the male and female gametes. The exaggerated plumage, coloration, behavior, and morphology of males are correlated with but not necessary to reproduction.
Both of Darwin's patterns are reflected in the language of natural history. In many species, the male is so conspicuously different from the female that the common name of a species describes only the male sex. Only male red-winged blackbirds (Aegelaius phonecius phonecius; Searcy 1979; Weather-head and Robertson 1979), are black with red epaulets on their wings, whereas the females are inconspicuous and dull brown in color (fig. 1.1a). In the bullfrog, Rana catesbania (Howard 1984), it is only the male that makes the deep call for which the species gets its common name; female bullfrogs are silent (fig. 1.1b). Only males carry balloons of silk as nuptial gifts for females in balloon flies, Hilara santor (fig. 1.1c; Kessel 1955). Neither the epaulets of the male blackbird, nor the call of the male bullfrog, nor the bower of the male bowerbird, nor the balloons of male balloon flies, are essential for sperm production or other physiological aspects of reproductive function. These males differ from the females of their own species in a rather arbitrary suite of phenotypic traits when considered across taxa.
Darwin used the term "trivial" to describe many of these exaggerated, male-limited characters because they appeared to have no clear relationship to viability or reproductive fitness. Despite very similar ways of life, closely related species could have males with very different phenotypes. Why should male tail length in one species be greatly elongated while, in another species of the same genus, males might possess a cape of expandable neck feathers and be rather ordinary in tail length (Gilliard 1962; Borgia 1986)? Why would longer tails be adaptive for males of one species but not the other? Furthermore, if these traits were adaptive for males, why were they not also adaptive for females? Darwin saw no obvious functional relationship between the exaggerated traits of males and the physical environment as he did for many other characters. Indeed, the "fit" between certain male phenotypes and the abiotic environment was exceptionally poor; exaggerated male characters might actually lower male viability.
The macroevolutionary pattern of large phenotypic differences between males of closely related species suggests that the selection responsible for these exaggerated male traits was rapid and strong. In contrast, we know from microevolutionary theory and empirical studies of artificial selection that selection acting on only one sex is considerably slower and weaker than selection acting in both sexes. In fact, selection on one sex but not on the other is only half as effective as selection acting on both, because half of the genes in any generation are derived from each sex in the previous generation (Falconer 1989). Selection restricted to one sex is tantamount to drawing half of the genes at random, and unselected. Selection that acts in opposing directions in the two sexes is slower still than selection absent in one sex. To understand the microevolutionary perspective on single-sex selection, consider an experiment in which a laboratory or captive population is subjected to artificial selection to increase tail length in males. There are several different ways that we might impose artificial selection and these have different effects on the expected rate of response. Consider first artificial selection on both males and females. After measuring tail lengths of all males, those with long tails are chosen as parents and those with shorter tails are discarded and prevented from breeding. We can quantify the strength of this selection using the standardized selection differential experienced by males, [S.sub.males] (fig. 1.2.a). The difference in average tail length between the selected males and the unselected males, divided by the standard deviation of male tail length, equals [S.sub.males]. Similarly, we measure tail lengths of all females and choose those with the longest tails for parents and discard those with shorter tails (fig. 1.2.b). The strength of this selection, [S.sub.females], is defined just like [S.sub.males], but relative to the female trait distribution. Total selection on our hypothetical population is the average of these two selection differentials, [S.sub.total] or ([S.sub.males] + [S.sub.females])/2. If our selection is as strong in males as it is in females, so that [S.sub.males] equals [S.sub.females], which equals S, then [S.sub.total] also equals S.
Now consider artificial selection only on males and not on females (fig. 1.3). As before, we measure tail length of males and select those with the longest tails as breeders (fig. 1.3a). However, we choose female parents at random with regard to tail length (fig. 1.3b). Thus, the selection differential in females, [S.sub.females], must be zero because the mean tail lengths of the breeding and nonbreeding females are the same. By using the selected males and unselected females as parents, fully half of the genes of each offspring, namely, those descending through the females, are not subject to any selection at all. Because half of the genetic material affecting tail length in the offspring generation has not been selectively screened but rather has been chosen at random, the total selection differential, averaged across the sexes, equals [S.sub.males]/2 or, if we select on males as strongly as we did above, S/2. This makes single-sex selection weaker by half than selection on both sexes. Single-sex artificial selection experiments conducted on a number of species confirm this theoretical expectation (Robertson 1980).
The taxonomic observation of conspicuous male divergence between closely related species stands in sharp contrast with the expectation from microevolutionary theory of a slower response to single-sex selection. This contrast between macroscopic pattern and microevolutionary process becomes even starker when we consider the evolution of sex-limited expression of male phenotypes. When genes are carried by both sexes but expressed in only one sex, we say that the expression of the gene is sex-limited. Most genes are expressed in both sexes; the limitation of a gene's expression to one sex is itself an evolved property of that gene or the developmental genetic system. Indeed, when artificial selection is practiced on only one sex, the focal character responds in both sexes, because selection in one sex does not limit the expression of the genes to that sex. For this reason, in our second artificial selection experiment (see above), we expect the tails of the unselected females to increase in length as a result of selection for increased male tail length. In the terminology of phenotypic selection models, the same trait (e.g., tail length) expressed in males and in females can be considered as two distinct but genetically correlated traits, one expressed in males and one in females. Because the genetic correlation across the sexes is positive, direct selection on one sex results in a correlated response in the homologous trait in the other owing to indirect selection.
With sexual selection, however, the selection differential in females is not zero, but actually less than zero (i.e., [S.sub.females] < 0). Why is this so? Because, as Wallace (1868) argued, the exaggerated traits favored in males are selected against in females. Thus, the sex-specific selection differentials are of opposite sign, [S.sub.females] < 0 < [S.sub.males]. It is as though we divided a population's genetic composition into two separate pools of genes: one from females and one from males. In our hypothetical population, in the male gene pool, we select for those genes that increase tail length or body size (0 < [S.sub.males]; fig. 1.4a). In the female pool, we select against these very same genes ([S.sub.females] < 0; fig. 1.4b)! We then combine the two divergently selected pools by mating the selected males and females to create the offspring. The result is an evolutionary process that is even weaker and slower than single-sex selection (fig. 1.3) because the average selection differential on the trait [([S.sub.males] - [S.sub.females])/2] is less than that of single-sex selection, [S.sub.males]/2. If the selection differentials were equal in the two sexes (but of opposite sign), then [S.sub.total] would equal zero.
How does sex-limited gene expression arise? The evolution of gene expression that exists in one sex and not the other is believed to occur via the evolution of modifiers, genetic factors that modify the normal pattern of gene expression during development (Fisher 1928; Altenberg and Feldman 1987). The fitness advantage to the modifier, which is otherwise neutral, accrues when there is selection favoring the phenotype in one sex and opposing it in the opposite sex. That is, the evolution of sex-limited gene expression requires a sex difference in the direction of selection, as occurs with sexually selected traits. As modifier genes spread through the population, the expression of the genes affecting tail length is diminished in the female sex and the trait can become sex-limited. Such modifiers can be viewed as genes that reduce the genetic correlation between a trait expressed in males and the same trait expressed in females. Reducing the genetic correlation reduces the limitation imposed by the sex difference in the direction of selection. However, the evolution of such modifier genes takes time, so the rate of evolutionary response for genes with initial phenotypic expression in both sexes is slowed.
There is an additional reason for sexual selection to be a slow evolutionary process. Exaggerated male traits do not appear to be adaptive at all life stages even in the male sex. Darwin (1859) reasoned that seasonal patterns in the expression of exaggerated male-limited traits indicate that these phenotypes may not be of selective advantage at other times during the life of males. Wallace (1868) went further, suggesting that such traits were selected against at these other times (see chapter 10). This conflict in the direction of selection at different life stages in the male sex makes total selection on these traits weaker and thus makes their evolution slower.
Consider again our population subject to artificial selection on males (fig. 1.5). In this population, suppose that selection is imposed on males at two different stages in the life cycle. When males are young and immature, selection may favor small tail size and act against males with larger tails, so that [S.sub.males] (early) < 0 (fig. 1.5a). Only males with the smallest tails at this stage are permitted to mature. Later, at maturity, selection among the remaining males favors only those with the largest tails becoming parents, so that [S.sub.males] (late) > 0 (fig. 1.5b). Genes that increase tail length at all ages will experience conflicting selection pressures, and, in particular, they will be culled and discarded at the first episode of selection. The total selection differential in males is the combination of these two opposing components, further weakening the overall strength of selection for these traits in the male sex.
Only those genes that fortuitously act later in male life to increase tail size will experience a coherent selection pressure and avoid the opposing juvenile selection. It is possible, and even likely, that there are "modifier" genes that could delay the timing of expression of tail size genes to late in male life. However, age-limited expression, like the sex-limited expression discussed above, represents another derived property of the genetic architecture and it further slows the expected rate of evolution under sexual selection.
Such conflicts between the fitness components of male viability and male reproduction may exist for many exaggerated male traits. It is often hypothesized that exaggerated male traits lower male viability by making males more conspicuous to predators (e.g., bird territorial calls, plumage, and displays) or by imposing high energetic costs on males (e.g., male vocalizations in frogs and toads). In some species, these costs to males have been well documented. For example, only male frogs call, and calling has been shown by ecological physiologists to be extremely costly in energetic terms (Ryan and Tuttle 1981). Males of many anuran species expend large amounts of energy calling and, although more than 90% of matings will take place over three or four nights, the males might call every night for two or three months. Also, in at least one species, Physalemus pustuluosus, males are victims of frog-eating bats that use the sex-specific calls to locate their prey (Ryan and Tuttle 1981; Ryan 1983b).
There is a similar risk to males in some species of lampyrid beetles or "lightning bugs" (Lloyd 1966).
Continues...
Excerpted from Mating Systems and Strategies by Stephen M. Shuster Michael J. Wade Copyright © 2003 by Princeton University Press. Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
