The problem of measuring trait-preference correlations without disrupting them

• Published in: Behavioral ecology 2019-11, Vol.30 (6), p.1518-1521
• Main Authors: Hosken, David J, Wilson, Alastair J
• Format: Article
• Language: English
• ISSN: 1045-2249; 1465-7279
• DOI: 10.1093/beheco/arz122

Abstract A key element at the heart of the Fisher–Lande–Kirkpatrick model of the sexual selection process is the genetic correlation between (male) trait and (female) preference. The strength of this association is critical in determining a population’s evolutionary trajectory, which is why estimating its magnitude is so important. In the Lande model, the trait-preference correlation is solely established and maintained by mate choice, and although it is unclear how exclusively mate choice does this in nature, the experimental designs typically employed to measure trait-preference genetic correlations could be systematically weakening estimates by not allowing free mate choice (similarly with husbandry practices). The precise impact of the problem is unknown, and possibly unknowable, but simple solutions can be applied to ensure the accuracy of trait-preference correlation estimates is maximized. Sexual selection through mate choice should result in the buildup of associations between the genes for mate preference and the genes that make individuals attractive. Testing for this association is therefore important in our understanding of how sexual selection works. However, as we explain here, the very methods we use to test for the preference-attractiveness association may make it harder to detect.