Proximate and Ultimate Origins of Female Preferences

Stalk-eyed males originate as an ornament of quality used by females

Few studies of organisms allow one to test all the ingredients necessary to discriminate between the competing hypotheses of runaway sexual selection and good genes models. One of the key pieces of evidence missing in all of the previous studies is the phylogenetic history of sexual selection. How do we know that a male trait under sexual selection in the present-day as an indicator of male quality evolved specifically for the purpose of an advertisement of quality? If a trait evolved for a specific function, then we can refer to the trait as an adaptation that solves a problem of sexual or natural selection. The origin of a trait makes individuals in the species better adapted to their environmental conditions. When we consider the evolutionary origin of a trait, we are delving into ultimate issues that define why a trait evolved. How females choose males and what sensory systems are used, is an issue of proximate mechanism.

We might search for the answer of a trait's origin in the fossil record to get at the evolutionary history of a group, but female choices do not fossilize and many male ornaments are far to delicate to leave a trace in the rocks. We need flesh, not just bones, to get at the dynamics of sexual selection and mate choice. In recent years, behaviorists have turned to a new branch of the comparative method called phylogenetic reconstruction (Brooks and McLennan 1991). While I cover the subject of phylogeny and history more thoroughly in an upcoming chapter (Chapter 17), I will go over some of the basics here. A phylogeny is a family tree of relationships that describes the degree to which modern day species are related. The phylogeny is our best guess as to which modern species are most closely related to one another, and which might be most distantly related. On a phylogeny, species that are separated by long branches are less closely related compared to species that are separated by short branches.

Figure 10.25. Courtship in stalk-eyed flies involves displays by males in a lek. Females tend to choose males with the largest stalks. Leks tend to form on the root hairs of plants. (From (Wilkinson and Reillo 1994).

For example, there are three species of flies that have eyes located on stalks that are quite closely related to each other in a phylogenetic sense (Wilkinson et al. 1998). One of the three, Cyrtodipsis quinqueguttata is more distantly related to the other two, C. whitei and C. dalmanni. In both C. whitei and C. dalmanni, males possess eyestalks that are far longer than the female's eyestalks. In third species, C. quinqueguttata, the eyestalks of males and females are the same length. A difference between the sexes or the degree of sexual dimorphism is our first clue that eyestalk length is a sexually selected trait in the males. Wilkinson and colleagues (1998) tested the proposition that males of the two sexually dimorphic species use the eyestalks to attract females. They conducted classic female choice experiments where the female chooses between a short and long males, and manipulative experiments where eyestalk length was varied by the experimenter. Both sexually dimorphic species showed dramatic female choice for long eyestalks, but the monomorphic species showed no female choice for variation in eyestalk length (Wilkinson et al 1998). Eyestalks are sexy in two species of stalk-eyed flies but not the third.

Figure 10.26. Phylogenetic relationships among species of Stalk-eyed flies in the genus Cyrtodipsis. Cyrtodipsis whitei and C. dalmanni share more DNA base pairs in common and probably derived more recently than C. quingueguttata. Both of the two closely related species have the sexually dimorphic male trait, exhibit female choice for long eye stalks, and also possess a selfish gene that causes sex ratio changes in progeny. The parasitic selfish gene is the change in enviroment leading to the origin of sexual seleciton. The monomorphic species exhibits none of the traits. A parsimonious reconstruction of trait evolution places all three changes at the same poin. C. whitei and C. dalmanni inherit the traits from a common ancestor (see text) (redrawn from Wilkinson et al 1998).

The phylogenetic question relates to the origin of the male trait, and the origin of the female choice. Are eyestalks a sexually selected trait that has evolved to solve a new problem faced by Cyrtodipsis whitei and C. dalmanni. If so, these two species should have arisen quite recently relative to the monomorphic species C. quinqueguttata. We can think of eyestalks and female choice as derived traits relative to the more ancestral monomorphic condition seen in C. quinqueguttata. The phylogeny for the three species indicates that the branch length for C. quinqueguttata reaches deeper into the past compared to the more recent origin of the sexually dimorphic species, C. whitei and C. dalmanni. We can map the evolutionary changes in eyestalk length and female choice onto the phylogenetic tree. The simplest hypothesis would be that the monomorphic condition is ancestral (quite a logical one, I might add), and that the sexually dimorphic condition arose once, when the hypothetical ancestor of the two sexually dimorphic species split off from the ancestral monomorphic species. This places the origin of the traits before the two species split off from one another, but after the two split off from the monomorphic species (Fig. 10.26).

Figure 10.27. Alternative reconstructions of trait evolution for the origin of long eye stalks (open box) and female choice (dash). While each of the two phylogenies is plausible, the reconstructions of trait origin (or loss) would require twice as many evolutionary steps as that found in Figure 10.26.

We have just reconstructed a plausible evolutionary history for female choice and the male trait using information from the distribution of traits among modern-day species, and the phylogenetic relationships among modern-day species. In our hypothesis of trait origin we assumed that the change between the ancestral monomorphic species and the two more derived dimorphic species only occurred once; this would be the simplest hypothesis. Using this assumption is also referred to as the principle of parsimony. Parsimony assumes that evolutionary change is slow enough that more complex hypotheses of trait origin are far less likely than the simplest hypothesis. For example, the two dimorphic species could have acquired the traits in two separate origins. Alternatively, all three species could have acquired the trait through a common ancestor, but the monomorphic species subsequently lost the dimorphic condition. The last two more complex hypotheses would require twice as many total evolutionary steps than the parsimonious hypothesis. In the absence of any other information, the parsimonious hypothesis is taken to be a 'null hypothesis' for the distribution of traits in the phylogeny.

With a working hypothesis in place of when male trait and female choice arose in the phylogeny of stalk-eyed flies, we move on to the next question: why did eye stalk length evolve in the two sexually dimorphic species? To address this question, we need to reconstruct the selective environment that might have led to the stalk-eyed trait being useful in males as an indicator trait for females. Again, there is no fossil record for reconstructing the conditions that led to sexual selection for long eye stalks. Fortunately, there is genetic evidence of a major change that would alter the mating environment of the flies (Fig. 10.24). Its time to synthesize some previously discussed issues of genic selection (Chapter 4) and sex ratio evolution (Chapter 9). The two sexually dimorphic species both possess a selfish gene on the sex-determining X-chromosome that for lack of a better word, kills off the Y during meiosis and replaces the missing chromosome with a copy of the X that carries the selfish gene. Like all selfish genes, all it does is over-reproduce itself during meiosis such that males who are infected with the element tend to produce more X-sperm that carry the element than Y-sperm (see Chapter 4, t-alleles in mice). This causes a male carrying the selfish gene to produce sperm that carry X-chromosomes. A female that mates with an 'infected' male would produce mainly daughters.

The selective consequences of this are straightforward, if we recall that the theory of Fisherian sex ratio favors a 50:50 sex ratio. Species "infected" with the selfish gene produce female biased sex ratios in both nature and laboratory cultures. Females that produce a biased sex-ratio of female offspring are at a striking disadvantage. Fisherian sex ratio theory indicates that a genotype which produces a 50:50 sex is evolutionarily stable (see Chapter 8), and the presence of sex ratio bias in stalk-eyed flies means that mothers with female-biased sex ratios are producing lots of female progeny that will have trouble finding mates. A female--biased population is susceptible to invasion by a female that can produce a male-bias. Accordingly, any gene that restores the biased sex ratio in males back to a 50:50 ratio of X:Y sperm would cause 'discriminiating' females to produce sons and thereby give their offspring a mating advantage relative to females that mate indiscriminately. Sons would have lots of females to mate with. Such a gene has evolved on the Y-chromosome and this Ym gene negates the effects of the selfish gene located on the X-chromosome, Xd, that distorts sex-ratio. Interestingly, the genes controlling long eye stalks in males is closely linked to the Ym . Females use the stalk-eyed trait in males as an indicator of their superior genetic background in that males with long eye stalks are more likely to bear the Ym gene, a gene that restores the sex ratio of their progeny back to a 50:50 sex ratio.

The natural history and genetics of the stalk-eyed flies is the first demonstration of a sexually-selected male trait evolving to serve as an indicator of a male's genetic quality. This story has all the elements that are required to support the theory of good genes. The correlation between female choice and the male trait is present in the sexually dimorphic species. When laboratory stocks of the fly are selected for longer or shorter stalks, female choice evolves in a correlated fashion (Wilkinson and Reillo, 1994). Moreover, the stalk-eyed male trait is genetically linked to a gene that rescues individuals from the effects of a selfish genetic element. Finally, eye stalks orginate as an adaption to indicated male quality.

Wilkinson et al (1998) have suggested that many male Y-linked traits such as guppy spots may be commonly used by females to indicate the presence of selfish genetic elements. Selfish genetic elements may be quite common in nature since distortions in sex ratio are found in guppies, mice (see t-allele example, (Lewontin 1962), and Drosophila (Atlan et al. 1997). Female choice for males may be driven by the ever present force of selfish genes, of which one large class of "sex-ratio-drivers" serves to distort the primary 50:50 sex ratio that is evolutionary stable in the long run. For example, male mice that are heterozygous for the t-allele (+|t, discussed in Chapter 2) likewise produce a distorted ratio of sperm bearing t-allele. Any female mouse with a heterozygous genotype (+|t) will produce sterile sons if they mate with a heterozygous male (t|t causes sterility in 1/4 of her sons). Accordingly, females with a +|t have evolved mate discrimination that allows them to avoid mating with heterozygous males in favor of wild type males (+|+) that are uninfected with the selfish t-allele. The case of infections by selfish genes and female discrimination of males that carry the bad gene (e.g., mice) or carry a good gene (e.g., stalk-eyed flies) is certainly not the only situations in which indicator genes might prove useful. However bizarre, this example serves to highlight the power of sexually selected processes to couple male traits with female choice in the face of a dramatic infection of the genome by a 'parasitic gene'.

Sensory Bias and Proximate Explanations for Phylogenetic Patterns

Recently, behaviorists have applied phylogenetic techniques to test the order in which female choice originates in relation to the origin of a male trait. The case of the stalk-eyed flies provides support for the idea that the male trait and female choice evolved nearly simultaneously in response to a change in mating environment; a biased sex ratio. A recent theory (Endler 1992; Ryan and Rand 1993; Ryan 1997) relates to sensory biases in the nervous or sensory system of females that pre-disposes them to pick males for some traits over other males that lack the trait. Females choose males not because they perceive them as sexy per se, but because they are "attracted to them". Such biases are present in the ancestral species and they remain latent in a population until a male evolves a mutation. Because a mutation in ornament "exploits" pre-existing sensory bias found in females, the theory of sensory bias is also referred to as sensory exploitation. Certain stimuli (e.g., colors, shapes, movement) may be useful in certain contexts (e.g., feeding and foraging) and the nervous system of females (and males) is honed by natural selection to be efficient at picking out food items from a world that is overly rich in extraneous stimuli. In a sense, these parts of the nervous and sensory system may be co-opted by sexual selection and a mutant male that displays a trait that triggers a heightened response in females may have an advantage. A male's signal may become fine-tuned such that it maximally stimulates the female sensory system.

Theories of sensory bias postulate that sexually selected male traits evolve in a species where females have a pre-existing phylogenetic bias for certain kinds of signals. Evidence favoring the idea of a sensory bias would place the origin of female choice as an ancestral condition (e.g., occurring earlier) relative to the origin of the male trait. In this chicken and egg argument, the evolution of female choice precedes the evolution of the male trait.

Figure 10.28. Hypothetical relationships between a male trait and female preference with regard to their distribution in the phylogeny. The phylogeny describes ancestor descendant relationships between the three extant (modern-day) species. The distribution of the male trait in extant species is denoted with a square. The presence of the female preference is denoted by a circle. Absence of the traits is denoted by a minus. Notice that the male trait does not occur in the more deeply branching outgroup in either clade. The outgroup species presumably reflects the ancestral condition of the trait. a) Hypothetical relationship between a male trait and female preference that would suggest that preference evolved prior to the male trait. b) Hypothetical relationship between a male trait and female preference in which the hypothesis of correlated evolution for the male trait and female preference cannot necessarily be ascribed to a pre-existing bias (from (Sinervo and Basolo 1996).

Alexandra Basolo looked at a large genus of fish, Xiphoporus, which have evolved elongated swords on their tail fins (Basolo 1990; Basolo 1995). The sword-tail is used as a sexual ornament. In species where males possess a sword, females prefer males with long swords (no surprise). A phylogeny of Xiphophorus indicates that many species have derived swords. One member of the genus, X. maculatus, has the "ancestral" condition and lacks a sword. Females in this species are quite content to mate with their swordless males; at least until Basolo tempted them with sworded males. Basolo gave females from this ancestral species a choice between males of their own species which lack a sword, or males of their own species with a surgically attached a sword. To control for the effect of surgery on the swimming ability of the male, which might interfere with his courtship, the first group of control males lacked a visible sword, but they did receive a clear sword tied onto the base of the tail fin. The sworded males received an opaque sword tied on to the base of the tail fin. She placed the two males in a pairwise choice trial and females overwhelming choose males that had a sword tied on!

Figure 10.29. Distribution of sworded males in the genus Xiphophorus and Priapella. Because the females in species whose males normally lack swords (denoted by *) prefer males with swords, Basolo inferred that the preference for swords arose as a pre-existing bias. The "pie-diagrams" below the tips of the tree reflect the best guess condition (probability of sword is proportional to area of black) for the male trait in hypothetical common ancestors that are located at the nodes joining the tips (from (Schluter et al. 1997)).

Further work on a more remote ancestral species Priapella olmacea. indicated that female preference evolved quite early in the history of this group of fish.Once again, females prefer their own males with swords tied on, even though males do not naturally possess swords. Basolo interpreted the female preference for swords in two swordless species in terms as an "ancestral" or pre-existing bias for sworded males in these fish. The idea also explains the widespread distribution of swords in other species of the genus.

Species in the hypothetical ancestor had a pre-existing bias for the sword, and when this trait showed up in some descendant, the sword spread through the population like wildfire. The exact reason for a pre-existing bias in the fish is unclear. Basolo has suggested that the sword may resemble the males gonopodium, which is a penis used to fertilize the eggs in the female's brood pouch. This group of fish has internal fertilization and brood eggs in a female brood pouch. Females will only mate with a male that displays readiness in the form of an extended gonopodium. Basolo hypothesizes that the sword provides a supernormal stimulus that heightens the female's readiness for copulations. Support for the sensory stimulation hypothesis, is provided by strength of female preference in species which possess swords. In these species, the females prefer supernormal stimuli over normal length swords. More work needs to done on the exact sensory mechanisms that predispose female swordtails to choose males with elongate tails.

The tungara frogs provide evidence that males co-opt specific mechanisms for sound reception in the female ear. Ryan and his colleagues have investigated a similar case of phylogenetic sensory bias in the tungara frog of Central and South America (Ryan and Rand 1993). This frog collects at ponds and uses a call to attract females. In a simple experiment, the composition of songs can be digitally altered on the computer and then played through speakers. Females readily respond to the songs being played from a speaker. Positive female choice was scored by movement of the female towards one speaker or the other. Males in two of the species, Physalaemus petersi and P. pustulosus, possess a more complex call consisting of two parts: an initial whine, followed by a chuck. Males in the other two species, P. coloradorum and P. pustulatus, do not possess a chuck at the end of the song, but only call with the whine component.

To generate the synthetic chucks in species with chuckless males, Ryan and his colleagues took the species typical whine and digitally mastered a chuck at the end of the song. Choice experiments indicate that females in all species prefer males that have a chuck added at the end of the song, regardless of whether or not the males of their species possess a chuck. Thus, available phylogenetic evidence suggests that the ancestral species consisted of females that had a preference for the chuck. The innovative chuck arose in one branch of the phylogeny, presumably by a mutation, and the ensuing chuck spread through the population because females already had a pre-existing preference for the chuck.

Figure 10.30. a) A sonogram of the mating call of male tungara frogs, Physalaemus spp., containing a whine (long horizontal lines) and two chucks (stacked lines at the end). b) The small panel for each species gives the amount of energy in the call over time. The high energy chuck is restricted to the end denoted by brackets. The distribution of female preference and male traits in the phylogeny of tungara frogs: P+ preference present, T+ male chuck present, and T- male chuck absent. Even though the males of the two species considered to be more ancestral, P. coloradorum and P. pustulatus, do not use a chuck in their calls, females of these species appear to have an ancestral bias or pre-existing predisposition for males with the chuck. Thus, the existence of a chuck in the derived calls of P. petersi and P. pustulosus are thought to have originated because the hypothetical common ancestor at the root of the tree had females with the pre-existing preference. (from (Kirkpatrick and Ryan 1991).

The sensory bias or pre-existing bias found in tungara frogs has a mechanistic basis in the vocal apparatus of the amphibian ear. The mechanics of the auditory apparatus of the amphibian ear also explains the nearly universal preference that females frogs have for large bodied males. As might be expected, larger males can produce lower-pitched calls than smaller males. These kinds of calls are much more effective at stimulating the female's ear. Specifically, the sound waves enter the female's ear and stimulate a cluster of receptors referred to as the basal papilla, which are sensitive to the range of sounds in lower frequencies. The females also possess an organ called the amphibian papilla which is responsible for fine scale discrimination of sound frequencies. The amphibian papilla is present in the ears of all species of tungara frogs, and the amphibian papilla is maximally stimulated by the frequencies produced by the chuck at the end of the call. Why this derived auditory structure is present in tungara frogs is not known, however, the evolution of the structure would have predisposed this group to the evolution of a male type that could exploit the pre-existing sensitivities of the female ear.