Traditionally, aggression has been considered an indicator of competitive ability, such that selection in competitive environments should favour aggressive individuals. However, variation in aggression persists in natural populations. An obvious question then becomes: what maintains variation in aggression? Aggression is considered both a plastic trait and a “static” one, as we observe both intra- and inter-individual variation. Plasticity in aggression occurs when individuals modify their behaviour according to the environment, and static differences occur when individuals differ consistently in their aggressive phenotypes across environments. In this thesis, I examined three components – variation in resource availability, variation in the social environment, and differences in aggressive phenotype – and their interactions to understand how variation in aggression persists across generations. The general experimental approach employed aggressive and less-aggressive strains of fruit flies (Drosophila melanogaster) mixed at different ratios and placed in environments of varying resource availability to understand how resource competition and the social environment influence the evolution of aggression. Interestingly, I found that, in a low resource environment, individual survival was greatest for the low-frequency strain, which is expected to lead to the maintenance of aggressive and less-aggressive phenotypes by negative frequency-dependent selection (NFDS), in Chapter 1. I then uncovered a novel behavioural mechanism which can drive NFDS through a combination of disruptive selection and social plasticity (Chapter 2). Lastly, I demonstrate how the social environment experienced during periods of food limitation carries over to future environments (Chapter 3). My thesis demonstrates that selection can act on a population of aggressive and less-aggressive phenotypes, such that aggressiveness is not always the most fit phenotype. Further, I show how social plasticity plays a critical role in determining the fitness of individuals, as well as how individuals will behave in the future. Taken together, I provide novel insight into why there is variation in aggression.