Only three years have passed since the first volume on environment for multi- agent systems was published. In the meantime, the notion of environment has become an active subject of research in multiagent systems and has stimulated researchers of various types of multiagent systems to exploit new opportunities for their agent systems. In the preface of the first volume, Parunak noted that the environment has been considered as a first-class member in the artificial life (AL) from the early days of this research line. During the three last years, an increasing number of multiagent system studies have successfully applied bio-inspired approaches where the environment plays a crucial role. Several papers in this volume gener- alize this idea to different kinds of agent systems, making it a multiagent system design principle: “taking away complexity from the agents’ internal structure and putting it in the environment enables designers to manage complexity better.” This principle has influenced designers of artificial societies that apply norms to regulate agents’ interactions up to engineers that have developed real-world traffic control applications. Interestingly, the environment is also the key aspect of innovative approaches that fruitfully reconcile the cognitive and the reactive point of views by having both kinds of agents cooperating through a delegate multiagent system soft- ware architecture. Also very interesting is the fact that classical environmental processes such as diffusion and evaporation are extended to process high-level information, introducing the notion of cognitive stigmergy. At the first E4MAS workshop in New York 2004, participating researchers were mainly interested in comparing and sharing their visions on this both fun- damental but still fuzzy concept of environment. Despite the fact that there was already a general agreement that environments were essential for multiagent sys- tems, the very question “what is the very nature of an environment?” has not yet been fully answered. The papers in this volume clearly illustrate how the environment can actively support the mediation of interactions among agents. The environment can be provided with different mechanisms and processes that enable agents to interact in original and unexpected ways, opening interesting research perspectives for the multiagent systems community. So, even if there still remain some differences on how researchers translate the concept of environment within their own agency perspective, the benefits of considering the environment as a concrete and ac- tive entity is becoming more and more obvious. The E4MAS series volumes are witnesses of this general awareness. Including both selected and revised papers from the third E4MAS workshop as well as invited papers, this volume shows the use of the environment at dif- ferent stages of the life cycle of multiagent systems. Without doubt, we are only at the beginning of exploiting the true potential that lies in the dynamics of the environment for multiagent systems. We hope that this volume is a stimulus for researchers to further explore this fascinating research line further. Acknowledgements. We are grateful to the PC members for their critical review work. We also thank Alexander Helleboogh, Kurt Schelfthout, Tom De Wolf, Koen Mertens, Nelis Bouck ́e, Robrecht Haesevoets, Bart Van Eylen and Tom Holvoet for their efforts for E4MAS. A special word of thanks to Tom De Wolf for managing the Web site.