Evolution of divergent female mating preference in response to experimental sexual selection

Abstract

Sexual selection is predicted to drive the coevolution of mating signals and preferences (mating traits) within populations, and could play a role in speciation if sexual isolation arises due to mating trait divergence between populations. However, few studies have demonstrated that differences in mating traits between populations result from sexual selection alone. Experimental evolution is a promising approach to directly examine the action of sexual selection on mating trait divergence among populations. We manipulated the opportunity for sexual selection (low vs. high) in populations of Drosophila pseudoobscura. Previous studies on these experimental populations have shown that sexual selection manipulation resulted in the divergence between sexual selection treatments of several courtship song parameters, including interpulse interval (IPI) which markedly influences male mating success. Here, we measure female preference for IPI using a playback design to test for preference divergence between the sexual selection treatments after 130 generations of experimental sexual selection. The results suggest that female preference has coevolved with male signal, in opposite directions between the sexual selection treatments, providing direct evidence of the ability of sexual selection to drive the divergent coevolution of mating traits between populations. We discuss the implications in the context sexual selection and speciation.

Keywords: Coevolution; Drosophila; courtship song; experimental evolution; population divergence; speciation.

Figure 1

Representation of the four artificial courtship songs synthesized. The figure represents the interval of time between two consecutive pulses of song (IPI) of the four songs. All song parameters values are given in Table1. To test for an effect of a further exaggeration of E-like IPI (an even shorter IPI), the difference between E-like and M-like IPI was subtracted to E-like IPI value, to create an EE-like IPI song. M is for monogamous, and E is for polyandrous.

Figure 2

Mating preference functions for IPI of selection line females (E and M) for mating latency (A) and mating probability (B). The two figures show that E and M females present opposite mating preference functions for IPI, for both mating latency and mating probability. The letters represent the fitted values estimated by the mixed-model associated with the four artificial songs, depending on female sexual selection treatment. M is for females from monogamous lines and E is for females from polyandrous lines. Ninety-five percent confidence intervals around each estimated value are represented.

Figure 3

Mating preference functions for IPI of ancestral females, for mating latency (A) and mating probability (B). The two figures show that ancestral females present a relatively flat mating preference function for IPI, for both mating latency and mating probability. The letters represent the fitted values estimated by the mixed-model associated with the four artificial songs. A is for ancestral females. Ninety-five percent confidence intervals around each estimated value are represented.