In this doctoral thesis, I use basic principles of natural selection to understand and predict interpersonal human behaviour. Specifically, I focus my research on the various strategies people employ to form and maintain social bonds. I begin this dissertation by showing that similar patterns of preferences for ability (i.e., traits that affect one’s capacity to provide benefits; e.g., wealth, intelligence, attractiveness) and willingness traits (i.e., traits that affect one’s likeliness to provide benefits; e.g., kindness, generosity) are found across four types of long-term relationships. I argue that this is because all relationships serve as a means of gaining valuable benefits through social exchange. In Chapter 2, I present a two-part investigation of the trade-offs people make in forming either narrow social networks with strong connections or broad social networks with weak connections. I show that preferences for the trade-off between network size and intimacy vary depending on the type of social interaction, and discuss the possibility of domain-specific (i.e., within the workplace, personal life, neighbourhood, family, etc.) preferences in the network size and intimacy trade-off. In Chapter 3, I assess the relative costs and benefits of helping kin over non-kin. Two main findings were obtained: (1) costlier help was found to be directed disproportionately towards kin over non-kin, and (2) status was a motivating factor for investment in kin and non-kin. In Chapter 4, I examine how two key aspects of interpersonal dynamics – (1) tracking: the degree to which people monitor the behaviours of others; and (2) tolerance: the degree to which one is lenient of temporary imbalances – are influenced by emotional closeness, changes in the availability of partners in one’s social environment, and stable preferences for the distribution of outcomes. My findings suggest that tracking and tolerance have separate adaptive functions. To conclude, I discuss the major contributions of my research, future directions of study, and real-life applications.