Abstract
Determining the direction and magnitude of mating preference is fundamental to many questions in evolutionary biology. Unlike the measurement of traits such as body size, the measurement of mating preferences is likely affected by experimental design. Scientists must choose both the behavioral assay in which to measure preference and the metrics that serve as a proxy for preference. The accuracy of these assays and metrics, however, is often unknown and seldom tested. Here, we compared the accuracy of 3 assays (dichotomous choice, audience assay, and no-choice assay) and 3 metrics (association time, courtship bouts, and number of eggs produced) in the bluefin killifish, which possesses strong, conspecific mating preferences when in sympatry with rainwater killifish. We consistently detected preferences in both males and females when using metrics associated with mating (i.e., courting bouts and number of eggs spawned). However, we failed to consistently detect preference when using time as a metric. We then used all 3 assays and metrics to test for cascade reinforcement. Cascade reinforcement predicts that enhanced behavioral isolation between sympatric species creates enhanced behavioral isolation among populations within species. We tested whether male and female bluefin killifish had heightened preference for mates from native over foreign populations. We consistently detected female preferences for native males, but did not detect male preferences for native females. Reproductive isolation values also reflect these preferences. Ultimately, we illustrated the importance of using multiple approaches to evaluate and legitimize measures of mating preference for males and females choosing among mates in different contexts.
