This study concerns itself with the relationship between game design and emergent social behaviour in massively multiplayer online games and virtual worlds. This thesis argues for a legitimisation of the study of ‘communities of play’, alongside communities perceived as more ‘serious’, such as communities of interest or practice. It also identifies six factors that contribute to emergent social behaviour and investigates the relationship between group and individual identity, and the emergent ways in which these arise from and intersect with the features and mechanics of the game worlds themselves. Methodology: Under the rubric of ‘design research’, this study was conducted as an ethnographic intervention, an anthropological investigation that deliberately privileged the online experience whilst acknowledging the performative nature of both game play and the research process itself. The research was informed by years of professional practical experience in game design and playtesting, as well as by qualitative methods derived from the fields of Anthropology, Sociology, Computermediated Communications and the emerging field of Game Studies. The process of conducting the eighteen-month ethnographic study followed the progress of a sub-set of members of the ‘Uru Diaspora,’ a group of 10,000 players who were made refugees when the massively multiplayer game ‘Uru: Ages Beyond Myst’ was closed in February of 2004. Uru refugees immigrated into other virtual worlds, using their features and capabilities to create ethnic communities that emulated the culture, artefacts and environments of the original Uru world. Over time, players developed ‘hybrid’ cultures, integrating the Uru culture with that of their new homes, and eventually creating entirely new Uru and Myst-inspired content. The outcome is the identification of six factors that serve as ‘engines for emergence’ and discusses their relationship to each other, to game design, and to emergent behaviour. These include: • Play Ecosystems: Fixed-Synthetic vs. Co-Created Worlds: Online games and virtual worlds exist along a spectrum, with environments entirely authored by the designer at one end, and those comprised primarily of player-created content and assets on the other, with a range of variations between. The type of world will impact the sort of emergent behaviour that occurs, and worlds that include player-created content will be more inclined to promote emergent behaviour. • Communities of Play: Distributed groups formed around play demonstrate distinct characteristics based on shared values and play styles. The study describes in detail one such play community, and analyses the ways in which its characteristic play styles drove its emergent behaviours. • The Social Construction of Avatar Identity: Individual avatar identity is constructed through an emergent process engaging social feedback. • Intersubjective Flow: A social reading of the psychological notion of ‘flow’ that describes the way in which flow dynamics occur in a social context through play. • Productive Play: Countering the traditional contention that play is inherently ‘unproductive’ as some scholars suggest, the thesis argues that play can be seen as a form of cultural production, as well as fulcrum for creative activity. • Porous Magic Circles and the ‘Ludisphere’: The magic circle, which bounds play activities, is more porous than game scholars had previously believed. The term ‘ludisphere' is used to describe the larger context of aggregated play space via the Internet. Also identified are leakages between ‘virtual worlds’ and ‘real life’. By identifying these factors and attempting to trace their roots in game design, the study aims to contribute a new approach to the making and analysis of user experience and creativity ‘in game’. The thesis posits that by achieving a deeper cultural understanding of the relationship between design and emergent behaviour, it is possible to make steps forward in the study of ‘emergence’ itself as a design material.