Elaborate sexual ornaments evolve because mate choice exerts strong sexual selection favouring individuals with high levels of ornament expression. Consequently, even at evolutionary equilibrium, life history theory predicts that ornamental traits should be under directional sexual selection that opposes contrasting selection to reduce the costs associated with their maintenance. Otherwise, the resources used to maintain ornaments should be used to improve other life history functions. Elaborate female ornaments have only evolved in a few species, despite females commonly experiencing strong sexual selection. One explanation for this rarity is that male preferences for female ornaments may be self-limiting: females with higher mating success become less attractive because of the lower paternity share they provide to mates with every additional sperm competitor. The unusual species in which female ornaments do occur can provide rare insight into how selection can favour the expression of expensive characters in females despite their costs. The main goal of my thesis was to determine how sexual selection acts on exaggerated sexual ornaments, and give new insight into how these ornaments may have evolved, in spite of the self-limiting nature of selection on male preferences. To determine the strength of sexual selection acting on female ornamentation in dance flies, we developed new microsatellite markers to assess polyandry rates by genotyping stored sperm in wild female dance flies. We first used polyandry rates to determine whether ornament expression was associated with higher mating success in female Rhamphomyia longicauda, a species that has evolved two distinct and exaggerated female ornaments. Contrary to our predictions, we found no evidence that females with larger ornaments enjoy higher mating success. We then compared polyandry rates in R. longicauda to those of two other species of dance fly, one (Empis aestiva) that has i independently evolved female ornaments on its legs, and another (E. tessellata) that does not possess any discernable female ornaments. We also estimated the opportunity for sexual selection, which we found to be similar and relatively low in all three species. Moreover, the standardized sexual selection gradients for ornaments were weak and non-significant in all three species. Females with more elaborate ornaments, in both within- and cross-species comparisons, therefore did not enjoy higher mating success. Overall, these results suggested that sexual selection operates rather differently in females compared to males, potentially explaining the general rarity of female ornaments. Our amplifications of stored sperm were able to reveal more than just mate numbers. We developed new methods to study patterns of sperm storage in wild female dance flies. We investigated how the skew in sperm genotypes from mixed sperm stores changed with varying levels of polyandry. Our data suggested that sperm stores were dominated by a single male in R. longicauda, and that the proportion of sperm contributed by this dominant male was largely independent of the number of rival males’ sperm present in the spermatheca. These results were consistent with the expectation of males using sperm ‘offence strategies’ in sperm competition and that the most successful male is likely to be the female’s last partner before oviposition. As a whole, my thesis contributed new molecular resources for an understudied and fascinating group of organisms. It exploited these new resources to provide the first estimates of lifetime mating success in several related species, and suggested that the general prediction that ornament expression should covary with sexual selection intensity does not seem to hold in this group. Instead, both the unusual prevalence of ii ornaments and the inconsistent evidence for sexual selection that sustains them in dance flies may owe their existence to the confluence of two important factors. First, the conditions under which sperm competition occurs: as last male precedence is likely, males are selected to prefer the most gravid females to secure a high fraction of her offspring’s paternity as they are unlikely to mate again before oviposition. Second, potent sexually antagonistic coevolution between hungry females and discerning males: females have evolved ornaments to disguise their stage of egg maturity to receive the benefits of nuptial gifts, while males face the challenge of distinguishing between gravidity and ornamentation in females.

The evolution of elaborate male ornaments via sexual selection is well-understood while the selective pressures acting on female ornaments remains unresolved. Female ornaments in species with strong sexual selection on the male homologue of the ornament were originally thought to result from an intersexual genetic correlation. My thesis explores the evolution of ornaments in females due to direct selection by developing theoretical models and examining the biology of a neotropical mosquito (Sabethes cyaneus) with sexually homologous ornaments coinciding with male-male competition. I began by exploring the morphology of the ornaments in both sexes of S. cyaneus. Sexual dimorphism in the size and shape of the ornaments was slight and both male and female ornaments showed classic hallmarks of sexually selected traits. I then tested for direct selection on S. cyaneus male and female ornaments via mutual mate choice. I found evidence of male, but surprisingly not female, preferences for ornaments. I then further considered the evolution of male mate choice in polygynous species. First, I investigated whether male investment in courtship by S. cyaneus may result in a lower operational sex ratio and thereby reduce the costs associated with male mate choice. Male courtship did pose a significant longevity cost to male S. cyaneus. Second, I explored the possibility that a female preference for male courtship effort may contribute to the benefits of male mate choice in a series of population genetic models. The spread of a male preference gene can be driven by female preferences for male courtship when males court preferred females more. Finally, I found that female S. cyaneus are not benefitting from signalling to increase their mating rate as they are monandrous. My thesis therefore challenges standing sexual selection theory and suggests that sexual selection on females may be more widespread than previously thought. / Felaktigt tryckt som Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 729