A trade-off between natural and sexual selection underlies diversification of a sexual signal

A longstanding hypothesis in evolutionary biology is that trade-offs between natural and sexual selection often underlie the diversification of sexual signals in the wild. A classic example of this "selection trade-off hypothesis" proposes that males evolve elaborate and conspicuous ornamentation in low-risk environments where female preferences dominate selection on sexual traits, but they evolve muted and relatively cryptic sexual traits in high-risk environments where selection from predators acts against conspicuous sexual traits and female preferences potentially weaken or reverse. However, little direct empirical evidence supports this notion. Using the model system of Bahamas mosquitofish (Gambusia hubbsi)—where males have recently evolved greater orange coloration in their dorsal fins in blue holes lacking predatory fish relative to populations with fish predators—we tested this hypothesis using fish replicas differing only in dorsal-fin color. Specifically, we employed plastic fish models in a combination of field and lab experiments to directly examine conspicuity to predators and female preferences for dorsal-fin color. We found that orange-shifted dorsal fins resembling the color exhibited in predator-free populations appeared more conspicuous to predatory bigmouth sleepers (Gobiomorus dormitor) that are evolutionarily naive to mosquitofish. Wild-caught female mosquitofish preferred the orange-shifted dorsal-fin model during dichotomous choice tests; evolutionary history with predators did not affect female preferences. Similar mate-choice trials with lab-born virgin females also found preferences for the orange-shifted dorsal-fin model and revealed significant genetic variation for female preferences. Our study provides direct empirical evidence documenting a trade-off between natural and sexual selection in a colorful sexual signal.