The Architecture of Phenotypes in a Naturally Hybridizing Complex of Xiphophorus Fishes

The origin and maintenance of phenotypic variation has generated considerable interest among students of functional morphology, sexual selection and behavioral ecology. In particular, hybridization has been suggested as a phenomenon which may generate novel phenotypic variation. In this dissertation I focus on the Xiphophorus birchmanni - X. malinche hybrid system to assess the role of hybridization in altering behavioral, morphological, sexual and non-sexual traits. I determine the relationship between the sword sexual ornament and body condition to support previous work which suggests that the sword is an inexpensive means to increase apparent size. My findings support the prediction that, while body size is condition-dependent, the sword is not. I show a trend toward hybrid populations displaying increased phenotypic variance and reduced phenotypic integration in sexual ornaments and body size. These findings provide evidence for a potential answer to a central question in the study of sexual selection, that of reduced genetic and phenotypic variance in sexual ornaments as the result of persistent direction selection generated by female choice. I take advantage of reduced phenotypic integration in hybrids allowing the evaluation of locomotor performance across a broad range of multivariate trait values. Sexual ornaments did not impair swimming performance per se. Rather, the sword negatively affected performance only when paired with a sub-optimal body shape. I evaluated how natural hybridization changes the relationship between boldness and anti-predator response. In poeciliid fishes, bold individuals have increased survival in the presence of predators. This non-intuitive observation may result from bold individuals being more likely to engage in anti-predator behaviors. Counter to my prediction, bold individuals were less likely to perform a fast-start response to a predator threat. This correlation was consistent among populations and species but was only significant in hybrids. My findings suggest that hybridization could influence correlations between behavioral traits in a manner similar to that documented for morphological traits