and preference are present in both sexes, but only the gene appropriate for
each sex is expressed. For simplicity’s sake, assume that the population is
haploid. After the first episode of mating, males with complex songs will
have garnered greater mating success. Furthermore, alleles that determine
complex song will be in linkage disequilibrium with alleles that determine
preference for complex song. As the frequency of complex song evolves in
the population due to preference for complex song, the preference itself
will “hitchhike” along with the complex song, and the preference will also
evolve to be in a higher frequency in the population. The stronger the
preference for complex song, the faster the rate of evolution of song and,
through the genetic correlation of song and preference, the faster the rate
of evolution of preference. The process can continue until the advantages
of the enhanced male trait, song complexity in this example, are offset by
the natural-selection costs of producing the trait, such as metabolic or predation costs, for example. At such a point, the forces of sexual selection and
natural selection will be balanced, and trait and preference will reach an
evolutionary equilibrium.
Although runaway sexual selection has been a popular hypothesis among
population geneticists for a number of years, it has been difficult to marshal
much empirical support for it (Andersson 1994; Ryan 1997). This might be
because it is a transient process; once signal and receiver reach an equilibrium point, it is difficult to determine how they got there. One approach to
testing the theory is to demonstrate the genetic correlation between signal
and receiver either by conducting selection experiments or comparing
populations of the same species that differ in trait and preference.
There are no good examples of studies in acoustic communication supporting the runaway hypothesis. But as a matter of illustration, consider an
elegant study of visual signaling in stalk-eyed flies. The eyes are located at
the end of long stalks in both sexes, but the stalks are much longer, and thus
the eye span much greater, in males than in females. Females prefer males
with greater eye spans. Wilkinson and Reillo (1994) conducted bidirectional
selection experiments on male eye span and determined whether there was
a correlated evolutionary response in female preference for eye span. After
13 generations, females from large eye-span lines and the unselected lines
both preferred males with larger eye spans. Females in the lines for which
short eye span was selected preferred males with shorter eye spans. Thus,
the female preference evolved even though it was not under direct selection but instead because it was genetically correlated to the signal.
6.2.2. Good Genes Selection
Good genes selection has been viewed as an intuitively appealing and more
utilitarian alternative to runaway selection. Under this scenario, females
attend to signal variation to assess a male’s genetic quality for survivorship.
But what would keep signals honest? Why could males not cheat and evolve
signals that falsely indicate high genetic quality? In some cases, signals
5. Selection on Signals
253
each sex is expressed. For simplicity’s sake, assume that the population is
haploid. After the first episode of mating, males with complex songs will
have garnered greater mating success. Furthermore, alleles that determine
complex song will be in linkage disequilibrium with alleles that determine
preference for complex song. As the frequency of complex song evolves in
the population due to preference for complex song, the preference itself
will “hitchhike” along with the complex song, and the preference will also
evolve to be in a higher frequency in the population. The stronger the
preference for complex song, the faster the rate of evolution of song and,
through the genetic correlation of song and preference, the faster the rate
of evolution of preference. The process can continue until the advantages
of the enhanced male trait, song complexity in this example, are offset by
the natural-selection costs of producing the trait, such as metabolic or predation costs, for example. At such a point, the forces of sexual selection and
natural selection will be balanced, and trait and preference will reach an
evolutionary equilibrium.
Although runaway sexual selection has been a popular hypothesis among
population geneticists for a number of years, it has been difficult to marshal
much empirical support for it (Andersson 1994; Ryan 1997). This might be
because it is a transient process; once signal and receiver reach an equilibrium point, it is difficult to determine how they got there. One approach to
testing the theory is to demonstrate the genetic correlation between signal
and receiver either by conducting selection experiments or comparing
populations of the same species that differ in trait and preference.
There are no good examples of studies in acoustic communication supporting the runaway hypothesis. But as a matter of illustration, consider an
elegant study of visual signaling in stalk-eyed flies. The eyes are located at
the end of long stalks in both sexes, but the stalks are much longer, and thus
the eye span much greater, in males than in females. Females prefer males
with greater eye spans. Wilkinson and Reillo (1994) conducted bidirectional
selection experiments on male eye span and determined whether there was
a correlated evolutionary response in female preference for eye span. After
13 generations, females from large eye-span lines and the unselected lines
both preferred males with larger eye spans. Females in the lines for which
short eye span was selected preferred males with shorter eye spans. Thus,
the female preference evolved even though it was not under direct selection but instead because it was genetically correlated to the signal.
6.2.2. Good Genes Selection
Good genes selection has been viewed as an intuitively appealing and more
utilitarian alternative to runaway selection. Under this scenario, females
attend to signal variation to assess a male’s genetic quality for survivorship.
But what would keep signals honest? Why could males not cheat and evolve
signals that falsely indicate high genetic quality? In some cases, signals
5. Selection on Signals
253