Beauty in the Eyes of the Beholders: Colour Vision and Mate Choice in the Family Poeciliidae

Sexual selection plays a major role in numerous aspects of evolution. Many models have attempted to explain how mate preferences evolve both across populations within a species and across species. ‘Sensory bias’ predicts that the traits involved in mate choice will co-evolve with the tuning of the sensory systems responsible for detecting such traits. The family Poeciliidae is a classic system for studies of mate choice and provides an excellent opportunity to examine the co-evolution of preference for colour traits and the sensory system detecting such traits: colour vision. In this dissertation, I present a body of work investigating how colour vision differs across species and populations, thus exploring the potential role sensory systems have in shaping mate preferences. To do this, I focus on the opsin genes, which play a predominant role in tuning the wavelength sensitivity of cone cells – the detectors for colour vision. I found the Long Wavelength Sensitive opsins (detecting red/orange colours) experience high rates of gene conversion due to their genomic architecture. The effects of conversion may be influenced by the importance of red/orange in mate choice decisions. While traditional models of duplication and divergence suggest sensory repertoire expansion occurs slowly, I found hybridization can expand sensory repertoires in one generation. I have termed this process: Hybrid Sensory Expansion. I then focus on one species to show that differences in visual tuning (gene expression and allele frequency) co-vary with mate preferences across populations in a manner that is consistent with the Sensory Exploitation (SE) model for the evolution of female mate preferences. However, I go on to find that closely related, highly sympatric species differ in colour vision more across populations than across species within populations on mainland South America. This suggests that while SE could explain differences in mate preference across populations, it may not scale up to explain species level differences as generally assumed. Taken together, these results show that the evolution of visual tuning may not always evolve through traditional mutation-selection models and that visual systems are far more variable across populations within species than generally assumed.