Understanding how genomic blueprints are translated into organismal phenotypes is a central area of study in evolutionary biology. Conversely, there is also evidence that phenotypes can influence patterns of genome evolution by modulating both adaptive and neutral evolutionary processes, but whether and how this process plays out remains unknown for most traits and taxa. This dissertation explores these topics in the highly social ants (family: Formicidae). In Chapter One, I perform a large-scale analysis of patterns of natural selection associated with two social traits, worker reproductive capacity and worker polymorphism, across twenty-two ant genomes which span a majority of ant diversity. I found that workers’ ability to reproduce increases the extent of relaxed selection genome-wide, while worker polymorphisms may reduce the prevalence of positive selection in the genome. In Chapter Two, I use developmental transcriptomics to identify candidate genes involved in the development of dimorphic workers in the Florida turtle ant, Cephalotes varians. I then explore the evolution of these candidate genes across the turtle ant genus, in both monomorphic and dimorphic species. I identify novel genes and signaling processes involved in the development of minor workers and soldiers, and find that selection appears to act primarily on genes related to maintenance of morph phenotypes in adult ants, rather than on genes involved in the development of those phenotypes in pupae. Finally, in Chapter Three I explore the evolutionary origins of the genes involved in morph differentiation using phylostratigraphic approaches. I find that the genes involved in morph differentiation are disproportionately of ancient origin, suggesting that turtle ants may re-use ancient toolkit genes in the production of their novel soldier morph.